Bioinformatics

Edited by
Soumi Dutta
Saikat Gochhait
Information Retrieval
in Bioinformatics
A Practical Approach
Information Retrieval in Bioinformatics
Soumi Dutta · Saikat Gochhait
Editors
Information Retrieval
in Bioinformatics
A Practical Approach
Editors
Soumi Dutta Saikat Gochhait
Institute of Engineering & Symbiosis Institute of Digital
Management and Telecom Management
Kolkata, West Bengal, India Symbiosis International (Deemed
University)
Pune, Maharashtra, India
ISBN 978-981-19-6505-0 ISBN 978-981-19-6506-7 (eBook)

https://doi.org/10.1007/978-981-19-6506-7
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer
Nature Singapore Pte Ltd. 2022
This work is subject to copyright. All rights are solely and exclusively licensed by the
Publisher, whether the whole or part of the material is concerned, specifically the rights
of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on
microfilms or in any other physical way, and transmission or information storage and
retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc.
in this publication does not imply, even in the absence of a specific statement, that such
names are exempt from the relevant protective laws and regulations and therefore free for
general use.
The publisher, the authors, and the editors are safe to assume that the advice and informa-
tion in this book are believed to be true and accurate at the date of publication. Neither
the publisher nor the authors or the editors give a warranty, expressed or implied, with
respect to the material contained herein or for any errors or omissions that may have been
made. The publisher remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.
This Palgrave Macmillan imprint is published by the registered company Springer Nature
Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore
189721, Singapore
Contents
1 Bioinformatics Overviews 1
Ritu Pasrija
2 Artificial Intelligence in Biological Sciences: A Brief
Overview 19
Uma Dutta, Nikhil Danny Babu, and Girish S. Setlur
3 A Review of Recent Advances in Translational
Bioinformatics and Systems Biomedicine 37
Chittaranjan Baruah, Bhabesh Deka, and Saurov Mahanta
4 Application of Bioinformatics in Agricultural
Pest Management: An Overview of the Evolving
Technologies 63
Bhabesh Deka, Azariah Babu, and Uma Dutta
5 Application of Bioinformatics in Health Care
and Medicine 83
P. Keerthana and Saikat Gochhait
6 Information Retrieval in Bioinformatics: State
of the Art and Challenges 101
Sunita, Sunny Sharma, Vijay Rana, and Vivek Kumar
v
vi CONTENTS
7 Hybrid Support Vector Machine with Grey Wolf

Optimization for Classifying Multivariate Data 111
M. Revathi and D. Ramyachitra
8 Bioinformatics and Its Application in Computing
Biological Data 133
Sonali Patil and Annika Durve Gupta
Index 155
Editors and Contributors
About the Editors
Dr. Soumi Dutta She did her B.Tech in Information Technology from
WBUT and M.Tech in Computer Science Engineering from WBUT with
securing 1st position (Gold medal). She has Ph.D. from Indian Institute
of Engineering Science and Technology (IIEST, Shibpur).
Dr. Saikat Gochhait is Post-Doctoral Fellow from University of

Extremadura, Spain. He has been awarded with MOFA Taiwan Fellow-
ship and Ministry of Health Russian Fellowship. He has more than 50
publications in journals indexed in Scopus, WoS, ABDC, Google Scholar,
etc. And 6 books published with Springer and IGI Global indexed in
Scopus. He features in the P-Rank: A Publication Ranking, HCERES
2021.
Contributors
Dr. Azariah Babu completed Ph.D. under the guidance of Prof. Dr. T.
N. Ananthakrishnan and done Post-Doctoral Research under the guid-
ance of Professor (Mrs.) Silvia Dorn at the SWISS Federal Institute of
Technology Zurich (ETH) in the Institute of Plant Sciences, Applied
Entomology Zurich, Switzerland. He has more than 28 years of research
experience in basic and applied aspects of Entomology. He has been
vii
viii EDITORS AND CONTRIBUTORS
engaged in various aspects of insect-host plant-natural enemy interactions

during the past 28 years.
Nikhil Danny Babu currently pursuing my Ph.D. in theoretical
Condensed Matter Physics at IIT Guwahati. He is interested in low-
dimensional quantum systems and bosonization.
Dr. Chittaranjan Baruah research is related to the fields of (a) Bioin-
formatics: Structural biology, proteomics, molecular phylogeny; (b)
Zoology: Fish Biology & Fish biotechnology, Nano-bio-pesticides; and
(c) Conservation biology: Turtles & Tortoises of Northeast India.
Dr. Bhabesh Deka (Senior Member, IEEE) is currently a Professor
with the Department of Electronics and Communication Engineering,
Tezpur University. He leads the Computer Vision and Image Processing
(CVIP) Laboratory, Department of ECE, Tezpur University. He is also
a Principal Investigator of two major research projects sponsored by
All India Council for Technical Education (AICTE) and Indian Space
Research Organisation (ISRO), Government of India. His research inter-
ests include image processing, particularly inverse ill-posed problems,
computer vision, compressive sensing MRI, and machine learning for
biomedical signal/image analysis.
Dr. Uma Dutta is affiliated to Department of Zoology, Cotton Univer-
sity. She is currently providing services as Associate Professor. She has
authored and co-authored multiple Cell & Molecular Biology.
Dr. Annika Durve Gupta is working with B.K Birla College (A) in the
Department of Biotechnology.
P. Keerthana lives in Chennai and started her career as Electronics and
Communication Engineer and currently pursuing a master’s in Analytics
and Finance from Symbiosis Institute of Digital and Telecom Manage-
ment. As an avid reader, her dream of writing started during her college
days and has published a paper on IoT-based medical record-tracking
systems for patients. Her strong fascination for the Internet of Things has
driven her interest in the emerging field of Translational Bioinformatics.
Vivek Kumar is associated as Research Fellow at GNA University,
Jalandhar, Punjab, India.
EDITORS AND CONTRIBUTORS ix
Dr. Saurov Mahanta is associated with National Institute of Electronics

and Information Technology, Guwahati with major research area on
Bioinformatics.
Dr. Ritu Pasrija currently works at the Department of Biochemistry,
Maharshi Dayanand University. She does research in Fungal Genetics,
Drug Resistance and Organelle Contacts.
Sonali Patil is associated as Assistant Professor at the Department of
Bioanalytical Sciences, B.K Birla College (Autonomous), Kalyan, Maha-
rashtra.
Dr. D. Ramyachitra is affiliated to Department of Computer Science,
Bharathiar University. He has published numerous publications in various
national and international peer-reviewed journals and presented scientific
papers across the world. Because of the active association with different
societies and academies as well as the contributions, she has been recog-
nized by the subject experts around the world. Her contributions are
appreciated by various reputed awards. Her clinical and scientific research
interests include Grid Computing, Data Mining, and Bioinformatics.
Dr. Vijay Rana is working with GNA University, Punjab in the Depart-
ment of Computer Science.
Dr. M. Revathi is working with Bharathiar University in the Department
of Biotechnology.
Dr. Girish S. Setlur works in the field of Theoretical Condensed
Matter Physics. He is interested in understanding and accounting for the
properties of everyday bulk materials from a knowledge of the funda-
mental constituents of the substance and the fundamental physical laws
governing those constituents. He is the inventor of a new technique
called ‘non-chiral bosonization’ which is uniquely suited to study strongly
inhomogeneous Luttinger liquids. He has also invented the notion of
a non-local particle-hole creation operator and showed that it may be
used to diagonalize interacting Fermi systems in any dimension. And
also interested in topological materials, specifically their nonlinear optical
properties.
x EDITORS AND CONTRIBUTORS
Dr. Sunny Sharma is working with University of Jammu in the Depart-

ment of Biotechnology.
Dr. Sunita is working with Arni University, Himachal Pradesh in the
Department of Computer Science.
List of Figures
Chapter 7
Fig. 1 Steps in a classifier model to perform data classification 113
Fig. 2 Comparison architecture of existing and SVM–GWO 114
Fig. 3 Linear SVM classifier with a decision plane 119
Fig. 4 Flowchart of hybrid SVM–GWO classifier 121
Fig. 5 SVM–GWO accuracy comparison for conventional
approaches 124
Fig. 6 SVM–GWO sensitivity comparison for conventional
approaches 125
Fig. 7 SVM–GWO specificity comparison for conventional
approaches 127
Fig. 8 SVM–GWO time period comparison for conventional
approaches 128
xi
List of Tables
Chapter 3
Table 1 Resources for translational bioinformatics that are open
to the public 51
Chapter 5
Table 1 Bioinformatic tools may be classified as follows: 85
Chapter 7
Table 1 SVM–GWO accuracy comparison for conventional
approaches 123
Table 2 SVM–GWO sensitivity comparison for conventional
approaches 125
Table 3 SVM–GWO specificity comparison for conventional
approaches 126
Table 4 SVM–GWO time period comparison for conventional
approaches 128
xiii
CHAPTER 1
Bioinformatics Overviews
Ritu Pasrija
1 Background
In the 1970s, a Dutch theoretical biologist Paulien Hogeweg along
with Ben Hesper, first coined the term bioinformatics. They were inter-
ested in accumulating information regarding biological systems. Their
observation was that in addition to biochemistry and biophysics, it is
worthwhile to recognise bioinformatics as a research area and has the
potential to become ‘biology of the future’. This became true as in
these last 50 years, development of bioinformatics has happened at a
very fast pace. Although for a particular interval, persons viewed bioinfor-
matics as the software tools advancement method to support, accumulate,
manoeuvre, and scrutinise biological information. Whilst this application
is indeed a significant one in bioinformatics, this field has much more
potential than that. Both ‘bioinformatics’ and ‘computational biology’
are instrumental in accumulating enormous information of several parts
of natural science. So, it is important to understand the difference in
these two terms. On the one hand, bioinformatics uses computer science,
R. Pasrija (B)
Department of Biochemistry, Maharshi Dayanand University, Rohtak, India
e-mail: [email protected]
© The Author(s), under exclusive license to Springer Nature 1

Singapore Pte Ltd. 2022
S. Dutta and S. Gochhait (eds.), Information Retrieval in Bioinformatics,
https://doi.org/10.1007/978-981-19-6506-7_1
2 R. PASRIJA
statistics to molecular biology and create computational & statistical tech-

niques, which help in examination and management of biological data.
On the other hand, computational biology uses computational simula-
tion mode, mathematical models, and fundamentals in computer science,
genetics, anatomy, biochemistry, and statistics among others. Amalgama-
tion of these two has led to a new term called ‘Systems biology’, which
combines organism-wide information of natural science for acquiring a
comprehensive perception of a biological entity, like a bacterium. This led
to creating synthetic genomes, and soon a synthesised cell would become
a reality. Nonetheless, to understand the feasibility of this fancy hypoth-
esis, it is important to revisit the key discoveries in biological sciences,
which would also help in understanding the history of development of
‘bioinformatics’ as a separate branch.
2 History
During the 1950s, DNA and computers were not the important tools
in research and in biochemistry, investigations were largely happening
on mechanistic enzymes model. Many scientists in fact thought that
proteins are the carriers of genetic information, as DNA seemed too
simple to carry genetic information, whereas protein show a large number
of alternatives and complexity.
This, the major turning point in bioinformatics has to be DNA being
regarded as the genetic material. The first evidence for this came from
experiments of Oswald Avery et al. (1944), who revealed that DNA regu-
lates the characters in organisms, instead of proteins. This group studied
the uptake of pure DNA from a virulent Streptococcus pneumoniae (S.
pneumonia) bacterial strain, which has smooth round colonies (named
S); which could bestow virulence to even a non-virulent strain (rough
colonies, R) (Avery et al., 1944). Subsequent work by Alfred Hershey
and Martha Chase (in 1952) validated these findings that DNA of bacte-
rial cells infected by bacteriophages can be transmitted to other bacterium
and alter the phenotype of recipient cell (Hershey & Chase, 1952). Later
in 1953, James Watson, Francis Crick, and Rosalind Elsie Franklin finally
proposed the double-helix structure of DNA (Franklin & Gosling, 1953;
Watson & Crick, 1953). Further, it took additional 13 years in inter-
preting the amino acid codon and 24 additional years in improving the
first DNA sequencing technique. Thus, 1970–1980 witnessed a paradigm
shift from protein to DNA analysis. In 1970, Saul B. Needleman and
1 BIOINFORMATICS OVERVIEWS 3
Christian D. Wunsch established their dynamic programming algorithm

for the alignment of pair-wise protein sequences. After another decade
of first multiple sequence alignment (MSA) algorithms was developed, its
application was applied to other biological sequences (DNA and RNA).
Therefore, development in DNA always lagged behind proteins. Compar-
ison wise, the amino acids alignment is based on-identical, similar and
dissimilar amino acids (based on their chemical properties), whereas only
identical matches and mismatches are considered for DNA and RNA.
The MSA and its use to sequence-structure-function relationship is so
common and useful that from the 1980s onwards, the term ‘bioin-
formatics’ is mostly used to refer to the computational methods for
genomic data analysis. This big transformation opened the possibility to
sequencing whole genomes. Biologist, Fred Sanger, undertook the first
genome sequencing in a bacterial virus, called bacteriophage φ × 174
(5368 base pairs). Later, Craig Venter in the 1980s sequenced the first
organism, a bacterium Haemophilus influenza. Ernst Haeckel, in 1866,
used DNA sequences in phylogenetic inference and reconstructed the
first molecular phylogenetic trees from amino acids linear arrangements in
proteins, which show the closeness and related ness among species during
evolution.
Similarly, many more genomes are already sequenced (more than
1000), including of human, called “The Human Genome Project
(HGP)”, which completed in April 2001 by two independent groups
(McPherson et al., 2001; Venter et al., 2001). These recent life science
data explosions—such as genotyping, transcriptomics, or proteomics—
also became possible with the availability of genomes and opened gates
for new studies. The gold mine of enormous data later became freely
available at European Molecular Biology Laboratory (EMBL) site (www.
ensembl.org). Along with this, various bioinformatics tools also became
available including on this site, like—BLAST, Ensembl, primer synthesis,
phylogeny that rely on the accessibility of the cyberspace. Investiga-
tions on genomic sequences information, including humans unlocked the
applied prospects like—discovery of drug and their targets, as well as indi-
vidualised therapy. Thus, biologists end up being increasingly dependent
on computational scripts, written in scripting languages, such as C, C++,
Shell, Python, R, and Ruby. Before, we look at the research challenges,
algorithms, big data, retrieval of information, and application of bioinfor-
matics; it is crucial to understand the basics of biology and bioinformatics
(Porter et al., 2021).
4 R. PASRIJA
Basics of Bioinformatics: We know that all surviving entities are

composed up of units called cells, which contain the genetic material
(nucleic acid), and passed from one generation to the next. Many living
systems are made up of only one cell (unicellular), and one cell is the
whole organism; whereas in developed species like plants and animals, a
life form has more than billions of cells. Apart from that, cells can be
of two major types: prokaryotic cells and eukaryotic cells. The prokary-
otic cells do not have nucleus and genetic material lies open inside the
cell, whereas in eukaryotic cells genetic material is present inside a struc-
ture, called nucleus. Prokaryotic cells are generally unicellular, whereas
eukaryotic cells can be either unicellular, like baker’s yeast Saccharomyces
cerevisiae (S.cerevisiae) or multicellular organisms, like humans.
The nucleic acid is of two kinds in nature: deoxyribonucleic acid
(DNA) and ribonucleic acid (RNA). Most organisms’ genetic material is
DNA, although few cells might have RNA, as in certain viruses. Both
forms of nucleic acid are polymers, assembly of repeated units, called
nucleotides. A nucleotide involves three segments: a base, a pentose
sugar (ribose), besides a phosphate group. These bases vary in different
nucleotides and are of four types: Guanine (G), Thymine (T), Adenine
(A) and Cytosine (C). For RNA, the bases are identical like DNA, except
T replaced with Uracil (U). RNA is a single helix, whereas DNA is a
double-helix molecule, in which bases lie parallel to each other and form
bond with each other, called hydrogen bond base pairing. During bond
formation, T always pairs with A base through two hydrogen bonds
(double bond) and C always pairs with G through three bonds (triple
bond). This removes steric hindrance and stabilises the structure. When
an RNA strand duo with another strand, the pairing followed is A = U
and C ≡ G. These nucleotides join one after another linearly and generate
polymer.
The Central Dogma: DNA is the controlling element, but needs
to pass on the information to a kind of RNA, termed as messenger
RNA (mRNA), through a process called transcription and subsequently
data in mRNA is passed via translation to proteins. Three DNA/RNA
nucleotides code for one amino acid, which polymerise in a linear fashion
to make proteins. Like UUU code for an amino acid, phenylalanine and
AUG code for methionine. There are 20 standard amino acids, which are
written as alphabets in capital. The sequence of three DNA nucleotides
(called codon) decides the amino acid incorporated. The nucleotides (U,
C, A and G) in random sequence of three can give 64 possible combina-

tions, thus one amino acid may be coded by more than one combination
and grouped together. Among these 64, three codons act as stop codon
(UAA, UGA and UAG), and once incorporated, they stop the translation,
as they do not code for any amino acid.
The genetic material has different genes on it, which control the
different characters in an organism, its fitness, etc. Thus, scientist devel-
oped fantasy for this molecule, as genetic material manipulation is
possible, and desired characters, fitness, chances of survival is achiev-
able. Similarly, it also opened avenues for gene manipulation to correct
any disease phenotype. All this became possible, as research has given
insight into function of the genes. This led to immense advancement
of techniques to study genetic material, termed ‘genomics’. Interest-
ingly, the different cells of an organism although have same DNA, but
do not express all the genes present on their genetic material. Thus,
all cells do express some common genes, called ‘house-keeping’ genes,
whereas some genes are transcribed to RNA and to translated proteins
in exclusive cells or time interval. Like in foetal stage, the haemoglobin
synthesised is different from that of an adult and are product of different
genes. Once born, the foetal haemoglobin gene is not transcribed any
more. These differential DNA expressions pattern led to different RNA
profile in various cells and is studied under ‘transcriptomics’. Similarly, the
functional molecules, proteins presence, and their measure are covered
in ‘proteomics’. Apart from that, three-dimensional (3D) modelling of
biomolecules and biological systems is also of interest to biological
scientist.
All these advancements led to generation of a large amount of biolog-
ical information among different organisms and species. Thus, biological
scientists need to utilise computational and analysis tools for acquiring,
understanding, interpreting and sharing of these data. These have led to
the ‘Bioinformatics’ we know today, and became essential in organising
information in modern biology, as well as treatment. Therefore, it is not
exaggerating to conclude that bioinformatics is a multidisciplinary field,
which connects biology with computer’s knowledge, mathematics, statis-
tics, and physics. This has led scientists in essentially acquiring a good
knowledge of molecular biology, along with computer science for analysis
of bioinformatics data.
The Human Genome: The name ‘genome’ exactly refers to total
genes or whole of the DNA’s content in an organism or a cell. A regular
6 R. PASRIJA
human cell encloses 23 couples of chromosomes (separate threads of

DNA). The human genome has 22 autosomes, and female additionally
has two copies of X chromosome (XX) and male has one X and one Y
chromosome (XY). These 23 pairs of DNA threads have approximately ∼
20,000–25,000 genes in the human genome, which are generally protein
coding, although sometimes only RNA is transcribed, which has regula-
tory role. Besides protein coding region, the controlling sequences are
also there in genome, which includes—promoters, introns, intergenic
(between-gene) regions, and repetitive sequences in the genome. On an
average, more than half of the human genome is transcribed and trans-
lated, though a very small quantity of them is managed as mRNAs and
study of all RNA transcripts is included under ‘transcriptomics’.
3 Perspectives of Computer Science

and Information Technology
As earlier stated, the sequencing techniques opened an era of loads of
data, called big data and its usefulness relies on programming and software
development, and building enormous datasets of biological information
in research (Gochhait et al., 2021). This information is much more than
easily and efficiently interpreted by a biology researcher. Further, different
investigators may decipher the data in dissimilar ways, and even the same
researcher may make varying explanations, resulting in erratic and non-
uniform data processing. Sometimes, after interpretation, the rationale
behind may be lost or only loosely remembered. Occasionally, a researcher
exits a study group; the technique used to understand data also goes
with them and vanishes. Lastly, the researcher’s interpretation may be
prejudiced towards getting a predetermined consequence. Thus, bioin-
formatics being a systematic study rule out all these anomalies in science
and research. The important basis of bioinformatics includes.
Algorithms: These are the rules followed in computations and done
on both Linux and Windows platform. Various tasks rely on particular
algorithms, which are critical in both examining and accurately handling
of the data. The biological scientists and bioinformaticians choose the
computer science processes for sequencing, gathering, unravelling biolog-
ical functions and relationships, which finally helps in interpretation of the
information. These include DNA, RNA and protein alignments (could
be local and global), gene prediction, phylogenetic tree construction
database similarity search, motif detection, Markov chains or informa-

tion entropy and sequence logos, molecular modelling, etc. The FASTA,
BLAST, Spectral Forecast, Objective Digital Stains (ODSs), self-sequence
alignment and Discrete Probability Detector (DPD) algorithm are few
examples. A bioinformatics can use resources freely available over the
internet and some are paid softwares. The literature references and various
databases (genome, sequence, function structure) of molecular biology
can be searched at PubMed, PubMed Central, NCBI, EBI, ExPASy,
RSCB. Some important tools in Bioinformatics are introduced below:
1. Recovery and exploration of sequence: Linear sequence of DNA,

RNA and proteins is used for sequence alignment and provides
lots of information. It depends on particular algorithm like FASTA
BLAST, CLUSTAL X/W, etc. These are helpful for identity, simi-
larity and dissimilarity in sequences (homology), phylogeny search
analysis and phylogeny tree construction for relatedness among
species.
FASTA is a text-based arrangement for representing either base

sequences or amino acid (protein) order, in which base or amino acids are
symbolised with single-alphabet codes and blanks in between alphabets
are not permitted. Matches are displayed in black and red are treated as
mismatches in nucleotide alignment. It was developed by David J. Lipman
and William R. Pearson in 1985 and is used in many programming
languages like Python, PERL, Ruby, etc. A multiple sequence FASTA
layout is obtainable by concatenating various single FASTA sequence in a
single file. The FASTA sequence first line starts with a ‘>’ (greater-than)
symbol as shown below for two genes: Green fluorescent protein (GFP)
from jellyfish and insulin from humans.
GFP gene sequence of Aequorea victoria in FASTA format
>AGTAAAGGAGAAGAACTTTTCACTGGAGTTGTGACAATTCTTGTTGAATTAGATGGTGAT
GTTAATGGTCACAAATTTTCTGTTAGTGGAGAGGGTGAAGGTGATGCAACATACGGAAAAC
TTACCCTTAAATTTATTTGTACTACTGGAAAACTACCTGTTCCCTGGCCAACACTTGTTAC
TACTTTGACTTATGGTGTTCAATGTTTTTCAAGATACCCAGATCACATGAAACGGCACGAC
TTTTTCAAGAGTGCAATGCCCGAAGGTTATGTACAAGAAAGAACTATTTTTTTCAAAGATG
ACGGTAACTACAAGACACGTGCTGAAGTTAAGTTTGAAGGTGATACCCTTGTTAATAGAAT
CGAGTTAAAAGGTATTGATTTTAAAGAAGATGGAAACATTCTTGGACACAAATTGGAATAC
AACTATAACTCACACAATGTATACATTATGGCAGACAAACAAAAGAATGGAATCAAAGTTA
8 R. PASRIJA
ACTTCAAAATTAGACACAACATTGAAGATGGAAGTGTTCAACTAGCAGACCATTATCAACA
AAATACTCCAATTGGCGATGGCCCTGTTCTTTTACCAGACAACCATTACCTGTCCACACAA
TCTGCTCTTTCTAAAGATCCCAACGAAAAGAGAGACCATATGGTGCTTCTTGAGTTTGTAA
CAGCTGCTGGTATTACACACGGTATGGATGAACTATACAAACACCATCACCATCACCATCA
CTAG
Humans Insulin gene sequence in FASTA format
>AGCCCTCCAGGACAGGCTGCATCAGAAGAGGCCATCAAGCAGGTCTGTTCCAAGGGCCT
TTGCGTCAGGTGGGCTCAGGATTCCAGGGTGGCTGGACCCCAGGCCCCAGCTCTGCAGCAGG
GAGGACGTGGCTGGGCTCGTGAAGCATGTGGGGGTGAGCCCAGGGGCCCCAAGGCAGGGCACC
TGGCCTTCAGCCTGCCTCAGCCCTGCCTGTCTCCCAGATCACTGTCCTTCTGCCATGGCCCTG
TGGATGCGCCTCCTGCCCCTGCTGGCGCTGCTGGCCCTCTGGGGACCTGACCCAGCCGCAGCC
TTTGTGAACCAACACCTGTGCGGCTCACACCTGGTGGAAGCTCTCTACCTAGTGTGCGGGGAA
CGAGGCTTCTTCTACACACCCAAGACCCGCCGGGAGGCAGAGGACCTGCAGGGTGAGCCAACT
GCCCATTGCTGCCCCTGGCCGCCCCCAGCCACCCCCTGCTCCTGGCGCTCCCACCCAGCATGG
GCAGAAGGGGGCAGGAGGCTGCCACCCAGCAGGGGGTCAGGTGCACTTTTTTAAAAAGAAGTT
CTCTTGGTCACGTCCTAAAAGTGACCAGCTCCCTGTGGCCCAGTCAGAATCTCAGCCTGAGGA
CGGTGTTGGCTTCGGCAGCCCCGAGATACATCAGAGGGTGGGCACGCTCCTCCCTCCACTCGC
CCCTCAAACAAATGCCCCGCAGCCCATTTCTCCACCCTCATTTGATGACCGCAGATTCAAGTG
TTTTGTTAAGTAAAGTCCTGGGTGACCTGGGGTCACAGGGTGCCCCACGCTGCCTGCCTCTGG
GCGAACACCCCATCACGCCCGGAGGAGGGCGTGGCTGCCTGCCTGAGTGGGCCAGACCCCTGT
CGCCAGGCCTCACGGCAGCTCCATAGTCAGGAGATGGGGAAGATGCTGGGGACAGGCCCTGGG
GAGAAGTACTGGGATCACCTGTTCAGGCTCCCACTGTGACCTGCCCCGGGGCGGGGGAAGGAG
GTGG
GACATGTGGGCGTTGGGGCCTGTAGGTCCACACCCAGTGTGGGTGACCCTCCCTCTAACCTGG
GTCCAGCCCGGCTGGAGATGGGTGGGAGTGCGACCTAGGGCTGGCGGGCAGGCGGGCACTGTG
TCTCCCTGACTGTGTCCTCCTGTGTCCCTCTGCCTCGCCGCTGTTCCGGAACCTGCTCTGCGC
GGCACGTCCTGGCAGTGGGGCAGGTGGAGCTGGGCGGGGGCCCTGGTGCAGGCAGCCTGCAGC
CCTTGGCCCTGGAGGGGTCCCTGCAGAAGCGTGGCATTGTGGAACAATGCTGTACCAGCATCT
GCTCCCTCTACCAGCTGGAGAACTACTGCAACTAGACGCAGCCCGCAGGCAGCCCCACACCC
GCCGCCTCCTGCACCGAGAGAGATGGAATAAAGCCCTTGAACCAGC
Similarly, protein sequence in FASTA format for both proteins can be

written and provided below—
GFP protein sequence from Aequorea victoria in FASTA format
>MSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVT
TFSYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIEL
KGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSVQLADHYQQNTPI
GDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITHGMDELYK
Insulin protein in Humans in FASTA format
>MALWMRLLPLLALLALWGPDPAAAFVNQHLCGSHLVEALYLVCGERGFFYTPKTRREAED
LQVGQVELGGGPGAGSLQPLALEGSLQKRGIVEQCCTSICSLYQLENYCN
BLAST: Its full form is Basic Local Alignment Search Tool and
this procedure catches the region of similarity between sequences. This
was first proposed in 1990 by David J. Lipman and his team, and
one of highly cited paper with more than 65,000 citations (Altschul
et al., 1990). This program can compare both nucleotide (n-BLAST)
and protein primary sequences (p-BLAST) in two different variants. The
evaluation of sequence is finally used to calculate the statistical signif-
icance of matches. It is freely available on internet at ‘https://blast.
ncbi.nlm.nih.gov/Blast.cgi’. It can be performed on various operating
systems like UNIX, Linux, Mac, and MS Windows and is written in C
and C+ language. BLAST deduce useful and evolutionary relationships
between linear arrangements of bases/amino acids, as well as help identify
members of gene families. Like haemoglobin gene sequence in humans
can be compared with that of mouse. The input sequence is generally
provided in FASTA or gene bank format; whereas the output formats,
include HTML, plain text, and XML. Besides online, the program is also
available in free and paid download versions (BLAST+). The megablast
and discontiguous megablast are other variants with separate applications.
Clustal W/X: is an algorithm for multiple sequence analysis (MSA) of
DNA or proteins. It produces meaningful multiple sequence alignment of
divergent species, and computes best matches for the chosen sequences
and line them up so that the resemblances and the disparities can be
understood by viewing the cladograms and phylograms, both of which
are part of the Clustal W algorithm.
2. Graph theory: Graph theory, also called as graph methodology,

takes the help of graphs for comparison. Graph is a set of vertices
(lines), connected by edges (exist between two vertices), which
could be directed (with arrow), undirected, weighed, etc. It is
useful in showing networks or flow of communication and simplifies
complex relationship.
10 R. PASRIJA
3. Artificial intelligence (AI): Here the bioinformatics functions

impersonate the brainpower of the human with computers. Its
usefulness is understandable by realising the need and importance
in whole genome sequencing, sequence reconstructions, and gene
finder. It also generates vital tools for data processing.
4. Data mining: Data mining is vital for extrapolation and meaningful
significant information can be extracted from huge datasets. The
multifaceted arrays of data can be used as input and utilise a variety
of arithmetical and numerical techniques to unearth surprising
inconsistencies, like grouping and bundling algorithms for any
process. An example includes gene annotation in whole genome
sequence, domain, and motif discovery. Similarly, mass spectroscopy
can classify proteins, although hindrance in data mining may come
from variances in complexity, scale, number and the lack of an
accepted ontology.
5. Soft computing: Many devices are upgraded for storing the
biomedical information; still computers have its significant role and
hold a special place with researchers and biologists. This has the
unique progress of expressing the data of the gene. In addition,
it also expresses the bioinformatics data. This evolves with neural
network model and artificial neural networks. Similarly, this is the
easiest and reliable method for analysing the process. The ultimate
factor of this method is genomic and proteomic applications. This
is useful for the scientists to do the experiments that result in a vast
amount of data.
6. Simulation and modelling: The term simulation refers to compu-
tation, especially for advanced algorithms and softwares, which
are used for curation and analysing the sequence, functions, and
structures. Computer simulation is consistent, accommodating, and
movable method to support the information. The properties of
biomolecules, their interaction like protein–ligand interaction, drug
target analysis, enzyme catalysed reactions and protein folding are
very well understood with simulation-based methods. Thus, the
working of these molecules mimics the actual physiological events.
These simulation studies rely on mathematics, physics, biophysics,
and chemistry. Like, quantum mechanics and molecular mechanics
(QM and MM) and lively mock-ups of proteins are amalgamated
with multiple-scale approaches like diffusion models with cellular
automata (CA) in brain tumour replication experiments. Another
example of antifungal protein chitosanase/glucanase, soil bacterium

Paenibacillus sp. shows binding with potential compounds at C-
terminus of proteins, which became possible with ligand dockings
and free energy estimates (affinity predictions).
Drug development has benefitted a lot from simulation studies. It is

well-known that drug discovery and progress require almost 10–15 years,
which is not even costly, but also labour intensive. The time and cost
input can be significantly reduced with use of computational tools. Drug
development is one of the foremost goals of bioinformatics and is popu-
larly known as Computer-Aided Drug Designing (CADD) (Dong &
Zheng, 2008; Macalino et al., 2015). It starts with in-silico structure-
activity relationships (SAR) studies, which include binding approxima-
tion of potential drug molecules on various target sites, which could
be: either enzymes, receptors, ion channels, or transporters, inside the
cellular membranes. Computers are used in depiction of shapes of refer-
ence molecules 2D (2-dimensional) and 3D (3-dimensional) structure,
followed by evaluation of their active pharmacophores (important groups
involved in binding) and strength of these interactions at target sites,
termed ‘molecular docking’ (Looger et al., 2003).
Different computational softwares used for docking are AutoDock 4,
FLIPDock, Vina, SwissDock, UCSF DOCK, FRED, EADock, SWISS
MODEL, LOMETS, PatchDock, HADDOCK 2.2, FIND SITE and
ClusPro, etc. (Grosdidier et al., 2011; Pagadala et al., 2017). Some of
these are paid; however, AutoDock 4 and FRED are freely available.
These SAR studies generate large amount of data and can be useful in
predicting a lead compound, which has the potential to develop as a drug
in future. The success stories include—Sotalol (brand name Betapace) is
used to treat a type of fast heartbeat called, sustained ventricular tachy-
cardia. It slows the heartbeat by acting on potassium channels (Brugada
et al., 1990). Similarly, Amlodipine (brand name Norvasc) is used for
high blood pressure and coronary artery diseases, by inhibiting calcium
ion influx across cell membranes (Calcium channel blocker) (Fares et al.,
2016). Similarly, Daliresp® (Roflumilast) is a phosphodiesterase (PDE)
4 inhibitor and is used for treating chronic bronchitis, psoriasis, and
neuroinflammation by reducing inflammation (Dong & Zheng, 2008).
Although this method has its own limitations, as side effects may not be
entirely be predicted and thus requires subsequent validation with actual
laboratory experiments and clinical trials.
12 R. PASRIJA
Image processing: It is essential and assists the biologists and scientists

to view their research. This method displays every stage of accomplish-
ments practically.
With the various steps/techniques involved in bioinformatics, we can
now proceed with application of bioinformatics.
4 Application of Bioinformatics
Genome Applications: It starts with DNA sequencing, genome assembly,
annotation of genes and prediction of gene function (based on the simi-
larity to known genes), sequence analysis for comparative exploration,
evolutionary studies, etc. Algorithms such as BLAST, Clustal W and
FASTA provide clarification in sequence investigation and examination.
This has benefitted the proteomics, transcriptomics, and metabolomics
studies. Similarly, functional genomics studies involve RNA-sequence
alignment and differential expression analysis. Comparative genomics and
computational evolutionary biology shed light on major events in evolu-
tion and divergence. Besides them primer designing, restriction enzymes
map analysis, RNA fold, dot plot are other genome-based applications.
In predicting protein structure: The RCSB PDB (Research Collabo-
ratory for Structural Bioinformatics Protein Data Bank) provided the first
open access digital platform for researchers and is available at ‘https://
www.rcsb.org/’. It supports retrieval of 3D-structure data of biological
molecules, including proteins and involves protein sequence retrieval,
followed by virtually establishing the similarities with sequences of known
structures, present in the PDB (homology modelling) (Berman et al.,
2000).
Biomedical: In the biomedical field, bioinformatics tools have an over-
powering effect on the understanding of genome, molecular medicine,
personalised medicine, and preventive medicine. Novel information on
the molecular mechanism of any ailment makes it easier to efficiently treat
and prevent the disease. This makes it easier to investigate genes straight-
forwardly associated with numerous diseases. For all ailments, alike drug
is given to patients, but different people have different genotype, so it
is important to consider the variations, even subtle but significant (called
single nucleotide polymorphism, SNP) in patients’ response to drug. It
is not exaggeration that if DNA profile of a patient is analysed, the
medication would be as unbeaten and efficient as possible, especially in
chemotherapy. Like Tamoxifen, (commercially known as Nolvadex), is
used in breast cancer treatment, which works by binding on oestrogen

receptor in breast tissue. This drug works when a person is positive for
oestrogen receptor (ER +ve) and is ineffective in patients lacking this
receptor (ER -ve). Similarly, polymorphisms (different forms) of gene
for human leukocyte antigen (HLA) alters the immune response and
metabolism of drugs. Many other drugs effectiveness has been found
related to the genotype and often studied under genome-wide association
studies (GWAS).
This proves that success rate in treatment depends on patient’s sensi-
tivity/insensitivity to particular drugs, which is sometime dependent on
variable forms of genes. This, jeopardy of failure during therapy can be
lessened by performing GWAS tests beforehand.
Human microbiome and metagenomics: The word ‘microbiome’
refers to the entire population of microbes that live inside any specific
ecosystem. The human microbiome refers to all the microorganisms living
inside human body, without causing the disease, including intestine/gut
and includes various bacteria and fungi. Their gene and environmental
interactions affect the human physiology. The variations in terms of
dissimilar bacterial populations in different humans create a ‘unique
microflora’, which has been studied in population groups. Like autistic
children, harbour significantly fewer types of gut bacteria than healthy
children do (e.g. Prevotella and Coprococcus species) (Kang et al., 2013).
P. copri is involved in breakdown of protein and carbohydrate foods.
Preventive Medicine and Gene Therapy: Preventive medicine focus
is on general health, well-being and simultaneously preventing diseases,
disability and death (Gochhait, & Omale, 2018). In some cases, the
patient’s body becomes a laboratory for drugs trial, as gene responsible
for disease is still not known, or it could be a multi-gene disorder. Our
genetic makeup, environmental conditions, disease agents, lifestyles and
genetic predisposition decide the probability of acquiring a disease in
future and many people die every year from preventable diseases. These
could be chronic respiratory diseases, cardiovascular disease, diabetes,
certain infectious diseases, etc. Here, genetic testing can also be done
to screen for mutations for disease-related gene polymorphism (single
nucleotide polymorphism or SNP) that cause genetic disorders or are
predisposed to certain diseases like certain type of cancers or diabetes.
Similarly, gene-editing tools, can be useful in correcting the genotype
and alleviating the disease phenotype. Although method is still under
research, optimistically it would soon become a reality.
14 R. PASRIJA
Forensic Analysis and Bioinformatics: Forensic analysis is largely

based on DNA-related data, which is also used for personal identi-
fication and relatedness. Genomic tests are extensively used as legal
evidences in paternity disputes, cadaver recognition, insurance business
frauds, and other crimes. This is the reason that many countries are
instituting forensic databanks of frequent/serial lawbreakers and crim-
inals. It is possible due to DNA sequencing, evidence orderliness, and
machine learning algorithms to establish a Probabilistic Graphical Model
(Bayesian networks) (Bianchi & Liò, 2007). The fingerprints, DNA
samples, retinal/iris scan (unique patterns of a person’s retina blood
vessels) and tongue prints are various methods, which are equally effec-
tive like signature verification, voice recognition, and face recognition
(Radhika et al., 2016). Their biometric databases and similarity searches
are possible due to bioinformatics only.
Microbial Genome and Climate change studies: This refers to
exploring the details of the genetic material of microbes and isolating
the genes, which give them an unparalleled ability to survive in extreme
conditions. This application can have ample implications in the improve-
ment of ecosystem, well-being, energy and industrial benefits. A well-
suited example is Pseudomonas putida, a bacterium with synthesised
genome, where DNA of four different species is combined to develop
a petroleum-degrading phenotype during notorious oil-spill in ocean in
1980, which successfully cleaned the hydrocarbons floating in water and
putting underwater animals at risk. In 1994, the US Department of
Energy initiated the Microbial Genome Project, for the sequencing of
microbes useful for environmental cleanup, energy production, industrial
treatment, and minimising toxic waste. Similarly, projects that are more
ambitious can be planned like global climate change, which is chiefly due
to increasing levels of carbon dioxide emissions and one way to reduce
atmospheric CO2 is possible by exploring the genomes of microorganisms
that uses CO2 for carbon.
Biotechnology: There are many uses of bioinformatics to fasten
research in the field of biology, like automated sequencing of genome,
gene mapping, protein configuration, organism identification, drug devel-
opment and vaccine design. Still, there is more scope as biotechnology
field is a side street. Few examples are discussed below-
Crop Improvement: It is agreed that the growth in population led
to worldwide temperature changes and resulted in declined crop yields.
Thus, a major challenge is that people should not die of starvation and
this is feasible with a sustainable agricultural production model. Here,

comparative genomics aid to understanding genes functions, across plant
species. In order to achieve high-quality crops in a short period, bioinfor-
matics databases are used to design new technologies and varieties with
better productivity.
Veterinary Science: Mere sufficient food consumption is not adequate
for people to survive and stay fit. Our health is dependent on nutrients up
take from food. At times, for consuming nutrients, people today depend
on livestock as well. A great success achieved in modifying the animal’s
genotype through bioinformatics, which reduced the risk of possible
infection and increasing production.
Innovation in examining the animal species helps in understanding
the system genetics of complex traits and provides accurate predic-
tion. Specifically, focuses was given on sequencing genome of animals
including—horse, cow, pigs, and sheep. This led to the development
in total production, as well as health of livestock. Moreover, bioinfor-
matics has helped researchers in discovering new tools for the discovery
of vaccine targets.
Canine cancer approximately affect one in every three dogs and exist
as one of the leading causes of death, despite advances in conventional
therapies. The cancer pathophysiology is linked with alteration of cellular
gene expression, and bioinformatics toolbox promises to put forward
innumerable insights into molecular mechanisms, diagnostics and novel
therapeutic interventions for cancer. In dogs, osteosarcoma (OSA) affect
the lives of 85–90% of affected, within two years of identification, despite
uncompromising surgery and chemotherapy. Here, the death is more
frequently due to metastasis and thus studying the transcriptome of
tumours in bones, scientists have identified a possible therapeutic target.
OSA cells express a protein, called her2/neu, on their cell membrane, and
metastatic cells express this protein at much higher levels than tumour
primary cells, and thus a target for immunotherapy. Scientists created
a vaccine, which improves the immune response with obliteration of
her2/neu OSA cells. In a first phase of clinical trial, this vaccine show
significant improvement over other treatment options.
Another example is equine metabolic syndrome (EMS), an endocrine
disorder in horses is linked with obesity, resistance to insulin. Studies show
that this syndrome is not only linked with carbohydrate intake, lack of
exercise, but there is also a susceptible genotype in horses.
16 R. PASRIJA
All these examples suggest the wide and ever-growing applications of

bioinformatics in biological science and medicine.
5 Conclusions and Future Prospects

Human evolution came a long way, and credit goes to advances in impor-
tant fields under biotechnology, as diagnostics, drug inventions, clinical
health, and agriculture have heightened our financial and social stan-
dards. Although a lot is achieved, but still bioinformatics can further
help biotechnology in reaching new heights, and helping humankind, but
we must set the ethical boundaries and inventions should happen within
moral limits. Thus, besides inventions, a string surveillance and regulatory
system is need of the hour. Industry practitioners and academicians have
to look forward towards the adaptation of AI (Artificial Intelligence) in
gaining momentum through big data analysis, machine learning, social
media analysis, algorithm decision-making, simulation modelling, and
other techniques that is used for bioinformatics visibility in the global
market (Varsha et al., 2021).
References
Altschul, S. F., Gish, W., Miller, W., Myers, E. W., & Lipman, D. J. (1990). Basic
local alignment search tool. Journal of Molecular Biology, 215(3), 403–410.
https://doi.org/10.1016/S0022-2836(05)80360-2
Avery, O. T., MacLeod, C. M., & McCarty, M. (1944). Studies on the chemical
nature of the substance inducing transformation of pneumococcal types. The
Journal of Experimental Medicine, 79(2), 137–158.
Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H.,
Shindyalov, I. N., & Bourne, P. E. (2000). The Protein Data Bank. Nucleic
Acids Research, 28(1), 235–242. https://doi.org/10.1093/nar/28.1.235
Bianchi, L., & Liò, P. (2007). Forensic DNA and bioinformatics. Briefings in
Bioinformatics, 8(2), 117–128. https://doi.org/10.1093/bib/bbm006
Brugada, P., Smeets, J. L., Brugada, J., & Farré, J. (1990). Mechanism of action
of sotalol in supraventricular arrhythmias. Cardiovascular Drugs and Therapy,
4(Suppl 3), 619–623. https://doi.org/10.1007/BF00357040
Dong, X., & Zheng, W. (2008). A new structure-based QSAR method affords
both descriptive and predictive models for phosphodiesterase-4 inhibitors.
Current Chemical Genomics, 2, 29–39. https://doi.org/10.2174/187539
7300802010029
Fares, H., DiNicolantonio, J. J., O’Keefe, J. H., & Lavie, C. J. (2016).

Amlodipine in hypertension: A first-line agent with efficacy for improving
blood pressure and patient outcomes. Open Heart, 3(2), e000473. https://
doi.org/10.1136/openhrt-2016-000473
Franklin, R. E., & Gosling, R. G. (1953). Molecular configuration in sodium
thymonucleate. Nature, 171(4356), 740–741. https://doi.org/10.1038/171
740a0
Gochhait, S., & Omale, D. (2018, June). Analytical solution to the Mathemat-
ical models of HIV/AIDS with control in a heterogeneous population using
homotopy perturbation method (HPM). Advances in Modelling and Analysis
A, 55(1). ISSN 1258-5769.
Gochhait, S., et al. (2021). Data interpretation and visualization of COVID-19
cases using R programming. Informatics in Medicine Unlocked, 26(6). ISSN:
0146-4116.
Grosdidier, A., Zoete, V., & Michielin, O. (2011). SwissDock, a protein-small
molecule docking web service based on EADock DSS. Nucleic Acids Research,
39(Web Server issue), W270–277. https://doi.org/10.1093/nar/gkr366
Hershey, A. D., & Chase, M. (1952). Independent functions of viral protein and
nucleic acid in growth of bacteriophage. The Journal of General Physiology,
36(1), 39–56.
Kang, D.-W., Park, J. G., Ilhan, Z. E., Wallstrom, G., LaBaer, J., Adams, J.
B., & Krajmalnik- Brown, R. (2013). Reduced incidence of Prevotella and
other fermenters in intestinal microflora of autistic children. PLoS ONE, 8(7),
e68322. https://doi.org/10.1371/journal.pone.0068322
Looger, L. L., Dwyer, M. A., Smith, J. J., & Hellinga, H. W. (2003). Compu-
tational design of receptor and sensor proteins with novel functions. Nature,
423(6936), 185–190. https://doi.org/10.1038/nature01556
Macalino, S. J. Y., Gosu, V., Hong, S., & Choi, S. (2015). Role of computer-
aided drug design in modern drug discovery. Archives of Pharmacal Research,
38(9), 1686–1701. https://doi.org/10.1007/s12272-015-0640-5
McPherson, J. D., Marra, M., Hillier, L., Waterston, R. H., Chinwalla, A., Wallis,
J., Sekhon, M., Wylie, K., Mardis, E. R., Wilson, R. K., Fulton, R., Kucaba, T.
A., Wagner- McPherson, C., Barbazuk, W. B., Gregory, S. G., Humphray, S.
J., French, L., Evans, R. S., Bethel, G., … Max-Planck-Institute for Molecular
Genetics. (2001). A physical map of the human genome. Nature, 409(6822),
934–941. https://doi.org/10.1038/35057157
Pagadala, N. S., Syed, K., & Tuszynski, J. (2017). Software for molecular
docking: A review. Biophysical Reviews, 9(2), 91–102. https://doi.org/10.
1007/s12551-016-0247-1
Porter, T. M., & Hajibabaei, M. (2021). Profile hidden Markov model sequence
analysis can help remove putative pseudogenes from DNA barcoding and
18 R. PASRIJA
metabarcoding datasets. BMC Bioinformatics, 22, 256. https://doi.org/10.

1186/s12859-021-04180-x
Radhika, T., Jeddy, N., & Nithya, S. (2016). Tongue prints: A novel biometric
and potential forensic tool. Journal of Forensic Dental Sciences, 8(3), 117–119.
https://doi.org/10.4103/0975-1475.195119
Sharma, A., Ghosh, D., Divekar, N., Gore, M., Gochhait, S., & Shireshi, S.
(2021). Comparing the socio-economic implications of the 1918 Spanish flu
and the COVID-19 pandemic in India: A systematic review of literature.
International Social Science Journal, 71, 23–36. https://doi.org/10.1111/
issj.12266
Varsha, P. S., Akter, S., Kumar, A., Gochhait, S., & Patagundi, B. (2021). The
impact of artificial intelligence on branding: A bibliometric analysis (1982–
2019). Journal of Global Information Management (JGIM), 29(4), 221–246.
https://doi.org/10.4018/JGIM.20210701.oa10
Venter, J. C., Adams, M. D., Myers, E. W., Li, P. W., Mural, R. J., Sutton, G.
G., Smith, H. O., Yandell, M., Evans, C. A., Holt, R. A., Gocayne, J. D.,
Amanatides, P., Ballew, R. M., Huson, D. H., Wortman, J. R., Zhang, Q.,
Kodira, C. D., Zheng, X. H., Chen, L., & Zhu, X. (2001). The sequence
of the human genome. Science, 291(5507), 1304–1351. https://doi.org/10.
1126/science.1058040
Watson, J. D., & Crick, F. H. (1953). Molecular structure of nucleic acids; a
structure for deoxyribose nucleic acid. Nature, 171(4356), 737–738. https://
doi.org/10.1038/171737a0
CHAPTER 2
Artificial Intelligence in Biological Sciences:

A Brief Overview
Uma Dutta, Nikhil Danny Babu, and Girish S. Setlur
1 Introduction
Artificial intelligence (AI) is the term used to broadly describe intelli-
gence demonstrated by machines. In the natural world, humans and other
animals display intelligent behavior in the sense of navigating their envi-
ronments and solving problems to achieve an end goal like finding a
U. Dutta (B)
Department of Zoology, Cotton University, Panbazaar, Guwahati, India
N. D. Babu · G. S. Setlur
Department of Physics, IIT Guwahati, Guwahati, India
G. S. Setlur
G. S. Setlur
Mehta Family School of Data Science and Artificial Intelligence, IIT Guwahati,
Guwahati, India

https://doi.org/10.1007/978-981-19-6506-7_2
20 U. DUTTA ET AL.
mate or hunting prey, etc. Even micro-organisms have been observed

to demonstrate a certain level of intelligence at the population level in
forming complex colonial structures and communicating using chemical
signals to overcome adverse situations. In general a system either living or
non-living can be said to possess intelligence if it shows complex adaptive
behavior targeted toward achieving an end goal. We do not consider our
computers that we use for almost everything nowadays to be intelligent
because computers simply execute a task given a set of instructions by an
intelligent user. But what if we teach computers to make decisions inter-
nally without a human user telling it step by step what is to be done. We
would be able to solve a host of complex problems that are simply too
tedious for the human mind.
Artificial intelligence entered mainstream academia in 1956 when John
McCarthy coined this term at a workshop in Dartmouth college. This
field was founded on the assumption that human intelligence can be
precisely quantified mathematically such that a computer can be made
to simulate it. This field arose out of a need to understand how the
human brain processes information. In the early days, several approaches
to AI were studied, for example, research in neurobiology and cyber-
netics as well as symbolic AI that considered the idea that intelligence
can be reduced to symbol manipulation. However, from the twenty-first
century onwards it is evident that machine learning has become the most
successful approach to AI research so much so that the term artificial
intelligence has become synonymous with machine learning using neural
networks.
With the advent of faster computers and increasing access to large
amount of data has enabled great advances in machine learning, and data
intensive deep learning methods now dominate AI research and appli-
cations. AI has come a long way since its inception and now there are
machine learning algorithms that classify images with greater than 97%
accuracy and there are AI programs that can beat the best human Go
player on the planet. AI-based models now influence almost every aspect
of our lives from online shopping, entertainment, to diagnosing illnesses.
In this chapter, we will discuss how AI/machine learning has influenced
modern medicine as well as research in the life sciences. In the following
sections, we shall briefly discuss some of the recent developments at the
interface of AI and life sciences. The objective of this chapter is mainly to
incite interest in the reader about this topic and not to serve as a technical
reference material. So keeping this in mind we brush under the carpet the
2 ARTIFICIAL INTELLIGENCE IN BIOLOGICAL SCIENCES: … 21
technical details and broadly discuss the state-of-the-art applications of

AI-based machine learning models in the health and life sciences.
2 Basics of Machine Learning (ML)

2.1 Building Blocks of a Neural Network
Machine learning is the most successful approach to AI. It is basically
a computer algorithm that learns through experience using data and
is usually optimized for a certain task, for example image recognition
or predicting customer behavior, etc. One analogy to this that can be
taken from daily life is that of a pair of new boots that feel uncom-
fortable and cause pain when worn initially but on repeated usage for
a couple of days the boots “learn” the shape of the wearers feet and
adapt for a comfortable fit. In machine learning, the computer program
is trained with sample data called ‘training data’ to make predictions
or decisions without being explicitly programmed to do so. Machine
learning is usually implemented using artificial neural networks (NNs).
Artificial neural networks were initially developed to mimic the biological
brain. A neural network consists of several artificial neurons and connec-
tions between these neurons. These artificial neurons are meant to mimic
a biological neuron. The construction of the network is such that the
neurons are arranged in layers. Every neural network should at least
have two layers, an input layer into which the input data is given and
an output layer that gives the final output of the NN. Additional layers
also called hidden layers can be included in between as needed and this
gives a “depth” to the neural network and the popular terminology ‘deep
learning’ is associated with neural networks that have several multiple
hidden layers. The hidden layers enable the NN to perform higher levels
of abstraction, for example in image recognition it is the hidden layers
that detect the edges and features in the image that helps in accurately
identifying the image. The number of neurons in a single layer is associ-
ated with the ‘width’ of the neural network. Enhancing the computational
efficiency of a neural network mainly involves finding a balance between
the width and depth of the network. Each neuron is a mathematical func-
tion called the activation function (f (z)) that takes an input value and
gives an output value. The activation functions used are nonlinear func-
tions. To simply put it, a neural network is a nonlinear function of many
variables that depends on many parameters as we shall see below. A linear
22 U. DUTTA ET AL.
function has the shape of a straight line. Any function that is not a simple
straight line when plotted is termed as a nonlinear function. It is neces-
sary to use nonlinear functions in the neural networks to get the higher
levels of abstraction needed for learning to take place.
The strength of each of the connections between the neurons in adja-
cent layers is quantified by assigning a number called a “weight” to each
of the connections. Now consider a neuron in one of the hidden layers.
The outputs of the neurons in the preceding layer are multiplied by their
assigned weights and all of this is summed up over the number of neurons
in that layer and an additional parameter called a “bias” is added to this
weighted sum, and this is fed as input to the activation function of the
neuron under consideration. This is how information flows from layer
to layer in the neural network. This process is called “feedforward” pass
through the network. So far we have encountered four parameters that
are important for constructing a neural network. We can think of these
parameters as knobs that one can adjust to design a neural network of the
desired size and configuration. Summarizing these parameters, they are:
Width: The number of neurons in each layer. This can vary from layer to
layer.
Depth: The number of hidden layers, i.e., the number of layers other than
the input and output layer.
Weights: The strengths of the connections between the neurons.
Bias: This is a parameter associated with each neuron. It is basically a
number that is added to the input of the neuron. It is also referred to as
an offset.
This encapsulates the basic structure of a neural network. Now we come

to the interesting part, how does the learning happen? How do we go
from here to accurately diagnosing Alzheimer’s from fMRI scan images?
How do we get it to make accurate predictions? This requires optimiza-
tion of the neural network and there is no given prescription for this, it
depends on the particular task at hand.
3 How Do Neural Networks Learn?

The human brain is truly incredible. A child who has seen a lion only in
pictures of a drawing book can immediately identify a real lion on a trip
to the zoo. Even such a simple task is very challenging for a computer
program. It requires a large dataset to train the neural network to produce

the correct output. Neural networks learn by training. The procedure
of training a neural network is what is known as machine learning. It
typically requires a huge amount of data to train a neural network and
obtaining the relevant data falls in the domain of data mining and data
science. There are three main techniques used in machine learning. These
are supervised learning, unsupervised learning, and reinforcement
learning.
Supervised learning: In this method, the model is trained with super-
vision analogous to a student learning in the presence of a teacher. The
input data is processed and labeled. Suppose we have a large collection of
images of animals. The task for the neural network is to correctly identify
the animal in each image. The input data is divided into a training set
and a test set. The data in the training set as well as the correct output
(names of the animals) will be supplied to the model and it will be trained
to correctly identify the animals in the training set. Once the training is
complete, the model will be tested by supplying images from the test set
and it will be able to correctly identify the images it has never seen before.
This technique is widely used for problems of classification and regression.
Unsupervised learning: In this method, the model infers patterns
from unlabeled input data. This method does not require supervision
and the algorithm discovers patterns in the data by itself. Suppose we
have a large collection of images of animals and we want to group the
images into different categories, like reptiles, birds, etc. This task can
be accomplished using unsupervised learning. This method enables the
neural network to identify hidden patterns in the data. This can be useful
when dealing with an unfamiliar dataset for which the labels, i.e., the
corresponding output data are not available. Unsupervised learning is
used for clustering and association problems.
Reinforcement learning: This is a feedback-based method of machine
learning in which the program learns from the outcome of its actions. It
receives positive feedback for correct actions and negative feedback for
incorrect actions. This method does not require any labeled data and is
used to design AI that needs to interact with its environment and navigate
through it like in the case of robots, game playing AI, self-driving cars,
etc.
We will briefly discuss only supervised learning as it is the one most
commonly used in several applications of AI in the biological sciences.
24 U. DUTTA ET AL.
The other two methods are beyond the scope of this chapter. In super-
vised learning, the labeled dataset is split into a training dataset and a
test dataset which is used to benchmark the trained model. During the
course of the training, the architecture of the neural network remains
the same, i.e., the depth and width parameters of the network are not
adjusted, but the weights and the biases (initially assigned randomly) will
keep changing until their optimum values are obtained. For this reason,
the weights and biases are called hyperparameters. In simple words, what
the neural network learns is the optimal values of all the weight and biases
that allow it to make accurate predictions. The concept of a cost function
is needed to understand this process. It is basically a mathematical func-
tion of the set of weights (w) and biases (b) denoted by C (w, b) and
it is proportional to the square of the difference between the correct
output/labels of the training data F (y train ) and the output produced
by the neural network F w,b (y train ). That is, C (w, b) ∝ F w,b (y train ) −
F (y train )2 . The idea is to find the set of weights and biases for which
the cost function C (w, b) is minimum. The cost function is like a penalty
for the NN and the target is to achieve the minimum possible penalty.
This is accomplished using “gradient descent” algorithm and backpropa-
gation. Gradient is the derivative of the cost function with respect to the
weight dC/dw. If the cost increases with increasing weight the gradient
will be positive, on the other hand, if the cost decreases with increasing
weight the gradient will be negative. The model needs to know whether
to increase or decrease the weights in order to minimize the cost func-
tion, the negative of the gradient (−dC/dw) shows exactly this. Now
the model knows in which direction to move the weights but we must
specify by what amount it must change the weights and this is decided
by the learning rate parameter η. The weights are updated according to
the following formula w = w − η|dC/dw| till the minimum of the cost
function (where dC/dw = 0) is reached. We have just briefly covered
the very basic introduction to gradient descent but in fact in practice
several gradient descent techniques are employed like ‘stochastic gradient
descent,’ ‘batch gradient descent,’ ‘mini batch gradient descent,’ etc. In
practice the situation is more complex than what we have described as the
cost function can have multiple minima and the gradient descent proce-
dure can get stuck around a local minima but the model should reach
the global minimum of the cost function in order for it to be successfully
optimized. This issue can arise if the initially assigned weights are near a
local minimum of the cost function. AI engineers use “backpropagation”
to ensure that the global minimum of the cost function is reached by the
gradient descent algorithm. The cost function is calculated at the output
layer and this information should be backpropagated to the previous
layers and all the weights associated with the neuron connections all the
way up to the input layer should be adjusted using the gradient descent
formula. This reverse flow of information in the neural network is termed
backpropagation. The mathematical details of this procedure are beyond
the scope of this chapter, and the interested reader can refer to any of the
excellent introductory books and articles on machine learning (Alpaydin,
2020; Baştanlar & Ozuysal, 2014; Kubat, 2017).
4 Applications of AI in the Life Sciences

With the rapid progress in data analytics techniques and computing power
artificial intelligence has become more accessible and it has made its
impact in a wide array of disciplines. In recent years, AI has made a
profound impact in health care, life sciences, and bioinformatics. Popular
AI tools such as deep learning neural networks are used in cancer,
neurology, and cardiology research. AI is no longer restricted to the
domain of computer science, and it has become paramount for researchers
and practitioners in other areas of science and arts to get exposed to this
field. In the following sections, we shall give a brief overview of some of
the recent applications of AI in biology and healthcare.
4.1 Using Deep Neural Networks to Understand the Biological

Brain
The biological brain is organized such that it has specialized areas for
performing different tasks. For example, there is an area in the brain that
recognizes objects but there is another region specialized for recognizing
faces in particular. Neuroscientists have struggled to understand why the
brain has specialized regions for various tasks. Now deep learning neural
networks are providing insights into why such specialization is so effective.
More and more neuroscientists are beginning to use AI to shed light on
the inner workings of the human brain. In image recognition traditional
neural networks as described in the previous section run into difficulties
when a shifted version of the input image is encountered. It does not
perform well if a picture of a cat curled into a ball is given to it and an
image of a cat fully stretched is given to it. It tends to overfit the data and
26 U. DUTTA ET AL.
hence underperforms when images of the same object in different posi-

tions is given to it. So the NNs are not translational invariant and are not
efficient in feature recognition. Also for high-resolution images traditional
NNs would require a large number of neurons and would require fitting
of an impractical number of weights and biases. A good AI model should
be able to identify a cat wherever it occurs in a picture and whatever posi-
tion the cat may be in, like running, sitting, or jumping. Convolutional
neural networks (CNNs) overcome all these problems.
In a convolutional neural network, filters are applied to the input image
in the hidden layers. For example, a filter that detects eyes could be
applied to the image and it detects how many times and in what loca-
tions an eye is present and creates a feature map. Similarly, filters that
detect other features are applied and feature maps are created. These
feature maps are then supplied as input to an activation function (usually
reLU). The output of the activation function determines whether a
certain feature is present at a location in the image or not. Adding more
layers of filters and feature maps will result in a deeper CNN that is
capable of detecting more abstract features. Pooling layers are used in
between convolutional layers, in which the largest values in the feature
maps are selected and passed onto the subsequent layers. This is called
max pooling and improves efficiency of feature detection. Finally, a fully
connected layer (all the neurons in the layer are connected to the previous
layer) performs the image classification in the end. A detailed explanation
of the CNN architecture can be found in (Albelwi & Mahmood, 2017).
In recent works by Bonnen et al. (2021) and Zhuang et al. (2021), the
authors have used a deep convolutional neural network (CNN) to clas-
sify images. They observed that the basic shape and outline of the objects
in the images were captured by the earlier stages of the neural network
while the more complex features were captured at the deeper stages in the
network. This is very similar to how the primate visual system works. In
the primate visual system, the pathway responsible for recognizing people,
places, and other objects is the ventral visual stream. The signals from the
eyes travel sequentially through the lateral geniculate nucleus to the V1,
V2, and V4 centers of the primary visual cortex and reach the inferior
temporal cortex. In the initial stages of the pathway, the basic features
like edges and shapes are detected and more complex features like eyes
and nose, for example, are detected in the later stages of the ventral
visual stream. The specificity in the functional match between the deep
convolutional neural net and the primate brain has gotten computational
neuroscientists excited about the prospect of using AI to understand

human brain function.
We do not have a good understanding of how the brain processes audi-
tory information. How the human brain processes sound in the auditory
cortex is still very much a open problem in neuroscience. In a recent
groundbreaking work by Kell et al. (2018) the authors designed a deep
neural network and optimized it to distinguish between two types of
sound: speech and music. Their goal was to find the best deep convo-
lutional neural net architecture that could perform this task efficiently
with minimum resources. The input given to the model was an audio clip
of speech with background noise or music with background noise. The
task given to the network was to identify the word spoken at a particular
time in the first case or to identify the genre of music playing in the latter
case. The researchers experimented with three possible architectures: (a)
Two separate networks specialized for each task with no shared layers.
(b) A network that is branched in the middle. Few of the initial layers
are shared. In the deeper part, the network is separated and higher-order
processing for each task occurs separately. (c) All the layers are shared
and process both the tasks, only the output layer is separate. The best
performing architecture was the one which was branched in the middle. It
was able to detect speech and recognize music efficiently using minimum
computing resources.
The separate networks with dedicated pathways for speech and music
tasks are the most accurate but it is also computationally expensive. The
network with separate speech and music pathways and a shared front end
was shown to be the optimum architecture, and it also agreed with predic-
tions that the deeper layers of auditory cortex have distinct regions for
processing speech and music. Moreover, this network made human-like
error patterns and predicted auditory cortical responses. The researchers
were able to infer the hierarchical organization in the human auditory
cortex from predictions made by a neural network. A deep neural network
was also successfully used to model the olfactory system in fruit flies which
is one of the most well-studied systems in neuroscience (Wang et al.,
2021). Artificial intelligence continues to make huge strides in helping
us understand better the inner workings of our own minds.
28 U. DUTTA ET AL.
4.2 AI in Medical Diagnosis

The incredible success of AI in image recognition means that it can be put
to use with confidence for important applications like medical diagnosis.
Early detection of diseases like cancer or Alzheimer’s disease could make
a huge difference in the outcome of the patient’s treatment. If AI can be
trained to predict the onslaught of cancer or early stages of Alzheimer’s
by looking at the CT scan or fMRI images before it becomes apparent to
a trained human doctor’s eyes it could go a long way in saving a patient’s
life. Several research groups all over the world are already tackling this
problem and huge strides have been made in the last five years. In the not
so far future AI medical diagnostics will be used to save thousands of lives
every year. But there are several challenges that need to be overcome first.
Convolutional neural networks (CNNs) are extremely adept at learning
patterns from data hence they are used extensively in chess playing AI,
computer vision, and even in language processing AI. But they require a
considerable amount of labeled training data and training time. There are
millions of labeled images of common objects and animals on the internet
but there is a shortage of correctly labeled and medically accurate training
data for diagnostic purposes. Another major limiting factor is that CNNs
are efficient only when the input data is two dimensional or planar. The
algorithm will run into errors if the objects are in 3D, for example a CT
scan image of body tissue.
A team of researchers have developed a framework for building neural
networks that can detect patterns on any kind of irregular surface. These
are called equivariant convolutional neural networks. These equivariant
CNNs can be used to learn the patterns in inhomogeneous curved
surfaces of body tissue like the heart, brain, lungs, and other organs.
Equivariance is similar to the idea of covariance in physics which means
that the physical laws do not change with change of reference frame. That
is, the laws of physics remain same for a stationary observer as well as
for a moving observer. Although the values of measurements made in
different frames can change, they can be transformed into each other
while preserving the relationship between different quantities. Equiv-
ariant CNNs are able to learn same patterns or features that occur in
different orientations or locations in the image even if the training exam-
ples don’t contain these different orientations thereby drastically reducing
the amount of training data required. This is highly useful for applications
like cancer detection as there is a shortage of labeled CT scan data in such
cases. Researchers have already obtained positive results in recognizing

lung tumors from CT scans using just a fraction of the training data used
in traditional NNs (Winkels & Cohen, 2019). The filter used in CNNs
is translational invariant in the sense that it learns a pattern in any loca-
tion of the image in 2D. In irregular curved surfaces the orientation (or
gauge) of the filter can change depending on the path it takes to traverse
the surface. This will affect the ability of the network to learn the features.
By fixing a certain orientation (gauge) of the filter the researchers found a
consistent way to transform every other possible orientation into it. Any
arbitrary gauge can be chosen initially but the transformation of other
gauges into the initial one should preserve the underlying feature. These
gauge equivariant CNNs were demonstrated to work on any arbitrary
surface (Cohen et al., 2019).
State-of-the-art 3D image processing CNNs were recently used to
learn features to diagnose Alzheimer’s disease which is a common
neurodegenerative disease (Huang et al., 2019). Alzheimer’s disease
mostly occurs in aged people and can cause severe cognitive impairment
and behavioral issues. The common biomarker indicating the onset of
Alzheimer’s is shrunken hippocampi. Usually by the time the diagnosis
is made using MRI scans, the patients already start showing symptoms.
Deep learning AI models can be trained using MRI images of early stage
Alzheimer’s to predict the onset of the disease before it is too late to
provide proper treatment to the patient.
4.3 Decoding Protein Structure Using AI

Proteins are large complex molecules that are fundamental to many
biological processes. Decoding the structure of proteins and how they
interact with each other is a challenging problem that has intrigued
computational biologists for several decades. The popular ‘protein folding
problem’ involves understanding the relationship between the protein’s
amino acid constituents and its final 3D structure. In the early days, super-
computers like IBM’s Blue Gene were put to this task and still it was a
cumbersome and time consuming effort. But in just a couple of decades
the rapid progress in AI research has culminated in programs like Google
DeepMind’s AI called AlphaFold that has made enormous progress
in determining a protein’s 3D shape from its amino acid sequence.
This has transformed bioinformatics and structural biology research.
AlphaFold’s predictions of some protein structures were identical to those
30 U. DUTTA ET AL.
obtained using experimental techniques like X-ray crystallography and

cryo-electron microscopy (cryo-EM). Most of the protein structures that
we already know were obtained using X-ray crystallography and now
cryo-EM is also being used. But this is a tedious process and it takes time
to arrive at the 3D protein structure. The use of AI like AlphaFold can
significantly speed up this process and can help in accelerating drug design
for diseases that results in faster production of vaccines and drugs for
new diseases. There are also other models like MaSIF (molecular surface
interaction fingerprinting) (Sverrisson et al., 2021) developed by inde-
pendent research groups that are less complex than AlphaFold but highly
successful in predicting protein structure. MaSIF targets the curved and
irregular 2D surface of the protein using geometric deep learning using
gauge equivariant CNNs which we have seen in the previous section. By
examining the surface of the protein, MaSIF can predict binding sites
on the surface. The researchers used surface features like surface curva-
ture, electric charge, and hydrophilicity to train the model. The model
then learnt to use these features to detect higher-order patterns. This AI
significantly speeds up the process if identifying potential fits among a
large set of proteins. MaSIF is not capable of predicting induced fit which
is when the molecular surface changes shape and chemistry in the vicinity
of another molecule to which it can potentially bind to. Researchers are
working on designing AI that can predict induced fit and other surface
dynamics as well. AI played an important role in our fight against the
COVID-19 pandemic that shook the world. In 2020 AlphaFold was
used to predict the structure of some SARS-CoV-2 proteins that were
not solved experimentally and its prediction of Orf3a protein in the virus
ended up matching perfectly with the cryo-EM structure. This played an
important role in synthesizing antiviral proteins to fight the virus. MaSIF
was used to predict the surface fingerprint of SARS-CoV-2 spike protein.
4.4 AI in Ecology and Conservation Biology

Every year a huge amount of ecological data is being collected by
researchers across the world. This includes satellite images, aerial images,
audio clips, camera traps, survey data, etc. This data can be used to
predict wildlife population movement, invasive species outbreak, changes
in vegetation and is even used in the fight against illegal wildlife traf-
ficking. Manually processing and analyzing these huge troves of data is
cumbersome and time consuming. Increasingly AI is being used for this
purpose and it can detect infrequent patterns and complex structures

in the data that is imperceptible to humans. This speedup achieved by
AI is necessary as researchers need immediate answers to pressing ques-
tions like whether conservation efforts are working. Computer vision
algorithms and audio processing algorithms both making use of CNN
architecture are commonly used in ecology and conservation. AI is more
accessible than ever and inexpensive ready to use AI applications are being
developed that can be used by people with no programming knowledge.
5 Limitations of AI
The field of artificial intelligence has come a long way since its inception
in the 1950s and now has become an indispensable part of technology.
But there is a caveat since we don’t yet understand completely the inner
workings of a deep learning model. There are several pitfalls and limita-
tions to AI that can easily be exploited to make it give completely wrong
predictions even for simple tasks. It seems that artificial intelligence is not
that intelligent after all and it should be used with caution with human
supervision. Blindly following the predictions of AI can be misleading in
some cases.
5.1 Overfitting
If a machine learning model gives correct predictions on training data
with good confidence but performs sub-optimally on new data used for
validation then it means that the model is overfitting the data. This
happens when the model starts learning the noisy features in the training
data in addition to the useful features. When there are more hyperpa-
rameters (weights and bias) to adjust than is necessary the model tends
to overfit the data and loses its predictive flexibility when supplied with
new data. To avoid overfitting the training process should be stopped just
before the cost function starts increasing after an initial decrease, if the
training process is continued for longer than necessary, then the model
becomes highly specialized to the training data won’t be generalized to
unknown data. Also selective dropping of some connections (weights)
during backpropagation will reduce the number of adjustable parameters
and decreases the complexity of the model to avoid overfitting.
Underfitting occurs when the model is not optimized properly and
performs poorly with the training data. To avoid underfitting more
32 U. DUTTA ET AL.
training examples should be included and the complexity of the model

also should be increased. A simple pictorial representation of underfitting,
overfitting, and balanced fitting of training data by a machine learning
model.
5.2 Adversarial Examples

Artificial intelligence can be tricked more easily than one could imagine.
A prime example of this is an adversarial example. An adversarial example
is input data that is modified to include some subtle noise that causes the
machine learning model to completely wrong predictions. This is best
illustrated in image recognition models where the corrupted input image
looks the same as the correct one to the human eye but the AI recog-
nizes it as a completely different object. This not only affects supervised
learning models but can affect reinforcement learning models too. This
can be used for malicious purposes as AI security systems can be thwarted
using adversarial examples and also self-driving cars can be sabotaged by
simply making minor changes to a stop sign such that the AI doesn’t
recognize it. A basic method to create adversarial examples is the ‘fast
gradient sign’ (Goodfellow et al., 2014) method in which the perturba-
tion or noise added to the input image (data) is N = E sign(∇ x C (w,
x, y)). Here ∇ operator means taking the derivative of a function with
respect to any of its parameters and the cost function C (w, x, y) depends
on the weights and biases denoted by w, x is the input to the model (pixel
values in case of images) and y is the target outputs. The gradient w.r.t
x tells us how the cost function changes with change in say the input
pixel values. What is needed is just the sign of the gradient as this tells
us whether the input pixel values should be increased or decreased. The
value E is taken to be very small so that the perturbation or noise added
is undetectable by the human users who input the data. The adversarial
input is the original input plus the added noise N .
Researchers have developed a few methods to defend against adver-
sarial attacks, one such method is ‘adversarial training.’ In this method,
adversarial examples are included in the training data with their corre-
sponding correct labels as targets and the numbers of training runs are
increased. But this takes up more resources and training time and it
is not foolproof against more sophisticated attacks. Another method is
‘Ensemble Adversarial Training’ which uses adversarial examples gener-
ated from several other pre-trained models to train the current model of
interest (Tramèr et al., 2020). This drastically reduces error rate due to
more sophisticated adversarial attacks.
5.3 Data Bias

Big tech companies pour in millions of dollars to develop more and
more sophisticated deep learning AI models but in the end its perfor-
mance heavily depends on the quality of data that is used to train the
model. Processing the data before feeding it to AI models is turning
out be a complex task. Skewed labeling of training data can render the
model inefficient in real-world situations even if it performs well in the
training process. This is particularly important when dealing with AI used
in medical diagnostics as biased or improperly labeled data can cause
misdiagnosis and could delay treatment procedures. One of the main
difficulties is in weeding out biases in the data that inadvertently creep
in due to human error. In a recent incident insufficient representation of
black Africans in training of Facebook’s image recognition, AI caused it
to group Africans and primates into the same group in an embarrassing
incident for the social network giant. This dependency on quality and
sometimes quantity of data is a limiting factor for the wide applicability
of AI.
6 Conclusions
In this chapter, we have given a brief overview of the impact of artifi-
cial intelligence (AI) in the biological sciences and bioinformatics (Varsha
et al., 2021). Using simple examples and basic terminology, we briefly
described the building blocks of AI and the steps that go into imple-
menting a successful model. Without burdening the reader with mathe-
matical detail we discussed what machine learning is and how the learning
process in a neural network works. We described how AI is being used to
push the frontiers in understanding the working of our brain and also how
it has become an indispensable tool in modern medical diagnosis. Some
of the most difficult scientific problems of the twentieth century like the
protein folding problem have become more tractable with cutting-edge
developments in AI in recent years. Exponential progress is being made in
this field day by day and it is becoming evident that artificial intelligence
together with human curiosity and innovation would be able to tackle the
biggest challenges that humankind is faced with.
34 U. DUTTA ET AL.
References
Albelwi, S., & Mahmood, A. (2017). A framework for designing the architectures
of deep convolutional neural networks. Entropy, 19(6). https://doi.org/10.
3390/e19060242. https://www.mdpi.com/1099-4300/19/6/242
Alpaydin, E. (2020). Introduction to machine learning. MIT press
Baştanlar, Y., & Ozuysal, M. (2014). Introduction to machine learning. In
MiRNomics: MicroRNA biology and computational analysis (pp 105–128).
Bonnen, T., Yamins, D. L., & Wagner, A. D. (2021). When the ventral
visual stream is not enough: A deep learning account of medial temporal
lobe involvement in perception. Neuron, 109(17), 2755–2766.e6. https://
doi.org/10.1016/j.neuron.2021.06.018. https://www.sciencedirect.com/sci
ence/article/pii/S0896627321004591
Cohen, T., Weiler, M., Kicanaoglu, B., & Welling, M. (2019). Gauge equiv-
ariant convolutional networks and the icosahedral CNN. In K. Chaudhuri,
R. Salakhutdinov (Eds.), Proceedings of the 36th International Conference on
Machine Learning, PMLR, Proceedings of Machine Learning Research (Vol.
97, pp. 1321–1330). https://proceedings.mlr.press/v97/cohen19d.html
Goodfellow, I. J., Shlens, J., & Szegedy, C. (2014). Explaining and harnessing
adversarial examples. arXiv preprint arXiv:1412.6572.
Huang, Y., Xu, J., Zhou, Y., Tong, T., Zhuang, X., & ADNI. (2019). Diagnosis
of Alzheimer’s disease via multi-modality 3d convolutional neural network.
Frontiers in Neuroscience, 13, 509. https://doi.org/10.3389/fnins.2019.
00509. https://www.frontiersin.org/article/10.3389/fnins.2019.00509
Kell, A. J., Yamins, D. L., Shook, E. N., Norman-Haignere, S. V., &
McDermott, J. H. (2018) A task-optimized neural network repli-
cates human auditory behavior, predicts brain responses, and reveals a
cortical processing hierarchy. Neuron, 98(3), 630–644.e16. https://doi.org/
10.1016/j.neuron.2018.03.044. https://www.sciencedirect.com/science/art
icle/pii/S0896627318302502
Kubat, M. (2017). An introduction to machine learning. Springer.
Sverrisson, F., Feydy, J., Correia, B. E., & Bronstein, M. M. (2021). Fast end-to-
end learning on protein surfaces. In Proceedings of the IEEE/CVF Conference
on Computer Vision and Pattern Recognition (CVPR) (pp. 15272–15281).
Tramèr F., Kurakin, A., Papernot, N., Goodfellow, I., Boneh, D., & McDaniel,
P. (2020). Ensemble adversarial training: Attacks and defenses. arXiv preprint
arXiv:1705.07204.
Varsha, P. S., Akter, S., Kumar, A., Gochhait, S., & Patagundi, B. (2021). The
impact of artificial intelligence on branding: A bibliometric analysis (1982–
2019). Journal of Global Information Management (JGIM), 29(4), 221–246.
https://doi.org/10.4018/JGIM.20210701.oa10
Wang, P. Y., Sun, Y., Axel, R., Abbott, L., & Yang, G. R. (2021). Evolving the
olfactory system with machine learning. bioRxiv. https://doi.org/10.1101/
2021.04.15.439917. https://www.biorxiv.org/content/early/2021/04/
16/2021.04.15.439917, https://www.biorxiv.org/content/early/2021/
04/16/2021.04.15.439917.full.pdf
Winkels, M., & Cohen, T. S. (2019). Pulmonary nodule detection in
CT scans with equivariant CNNs. Medical Image Analysis, 55, 15–26.
https://doi.org/10.1016/j.media.2019.03.010. https://www.sciencedirect.
com/science/article/pii/S136184151830608X
Zhuang, C., Yan, S., Nayebi, A., Schrimpf, M., Frank, M. C., DiCarlo, J.
J., & Yamins, D. L. K. (2021). Unsupervised neural network models of
the ventral visual stream. Proceedings of the National Academy of Sciences,
118(3). https://doi.org/10.1073/pnas.2014196118. https://www.pnas.
org/content/118/3/e2014196118. https://www.pnas.org/content/118/
3/e2014196118.full.pdf
CHAPTER 3
A Review of Recent Advances

in Translational Bioinformatics and Systems
Biomedicine
Chittaranjan Baruah, Bhabesh Deka, and Saurov Mahanta
1 Introduction
Translational research utilises scientific findings produced in the lab,
clinic, or field and turns them into novel therapies and medical care
methods that directly enhance human health. Translational research
aims to transfer fundamental scientific discoveries into application more
C. Baruah
Bioinformatics Laboratory, Postgraduate Department of Zoology, Darrang
College, Tezpur, India
B. Deka (B)
North Bengal Regional R and D Centre, Tea Research Association, Jalpaiguri,
India
S. Mahanta
National Institute of Electronics and Information Technology (NIELIT),
Guwahati, India

https://doi.org/10.1007/978-981-19-6506-7_3
38 C. BARUAH ET AL.
rapidly and efficiently. When it comes to productivity and translating

research into new healthcare advancements, it provides a wide range of
specialised resources. Research that fosters and supports multidisciplinary
collaboration between laboratory and clinical researchers considers the
requirements of communities, and discovers and promotes the adop-
tion of best medical and healthcare practices are all part of translational
research (Gochhait et al., 2021).
According to the stage in the translation process (from the commence-
ment of research to social application and effect), translational research
is categorised. The T-Spectrum (Translational Spectrum) below depicts
the many stages of translational research and development. Transcription,
translation, and mRNA and protein turnover are all part of gene expres-
sion. What happens when this dependence breaks down? (Buccitelli &
Selbach, 2020). A post-translational modification (PTM) affects the fate
of proteins in eukaryotic cells. Many web-tool predictors for different
PTMs are launched to help diagnose and prevent illnesses (Mohabatkar
et al., 2017).
Information technology and databases are used in bioinformatics
research to solve biological questions. Genomic and proteomic bioinfor-
matics applications are of global importance. The study of genomes is
known as genomics or genome research. A whole genome is a collec-
tion of DNA sequences that contain the genetic information that has
been passed down from one generation to the next over the ages.
It is a physical and functional unit of heredity that is passed from
parents to offspring via genetic inheritance. To summarise: Genomic
research includes the sequencing and analysis of all genetic material in an
organism—from genes to transcripts. To put it another way: Proteomics
studies all proteins together called the proteome. Beyond genomics and
proteomics, bioinformatics is utilised in a wide range of biological disci-
plines (i.e. metabolomics, transcriptomics). Bioinformatics is a branch
of study that aims to understand complex biological processes via the
use of computers. Protein structural alterations, activities, and functions
are regulated by post-translational modifications (PTMs) in almost every
biological process and activity. Understanding cellular and molecular
processes begin with protein PTM identification. However, unlike tedious
trials, PTM prediction utilising different bioinformatics methods may
offer accurate, convenient, and efficient techniques while also providing
important information for future studies (Liu et al., 2015). The introduc-
tion of next-generation sequencing (NGS) technology has accelerated the
3 A REVIEW OF RECENT ADVANCES IN TRANSLATIONAL … 39
identification of prostate cancer biomarkers Prostate cancer diagnosis and

prognosis remains difficult despite the deluge of sequencing data. Chen
et al. discussed high-throughput sequencing’s recent advances in prostate
cancer biomarkers (Chen et al., 2013a).
Many scientists now refer to systems biology as the next wave in bioin-
formatics. Integrating genetic, proteomic, and bioinformatics data creates
a holistic perspective of a biological entity. Systems biology may study how
a signalling pathway operates in a cell. Systems biology can simulate the
genes involved in the process, their interactions, and how changes affect
downstream consequences. Bioinformatics may be used in any system that
can express information digitally. From single cells to entire ecosystems,
bioinformatics may be used. The full “parts lists” of a genome help scien-
tists better understand complicated biological processes. Identifying how
these components interact in a genome or proteome is the subsequent
stage of intricacy in the research process.
2 Translational Bioinformatics
Translational bioinformatics (TBI) is a new area that applies biological
research to patient care and medication discovery. Develop and analyse
clinical and biological data to study illness heterogeneity using computer
methods. The search for disease gene(s) requires a thorough under-
standing of the complex network of biological mechanisms involved
in disease progression. This chapter aims to outline the biological and
clinical data integration strategy. It also explains the key datasets and
techniques used in translational bioinformatics to treat illnesses.
Translational bioinformatics focuses on utilising current research to
connect biological data with clinical informatics. Translational bioinfor-
matics now covers the biological and healthcare industries, bridging
the gaps between the bioinformatics and medical informatics. Transla-
tional bioinformatics has made several databases available to researchers.
These databases are useful for physicians, biologists, clinical researchers,
bioinformaticians, and health care researchers. These databases help biol-
ogists comprehend illness management and medication development
techniques, which help them, generate novel hypotheses. Gene variations,
enzymes, and descriptive genomics databases are examples of translational
bioinformatics databases.
40 C. BARUAH ET AL.
2.1 Categorising Translational Bioinformatics Research

Translational bioinformatics research can be basically sorted into four
categories (Denny, 2014):
● The utilisation of Clinical “big data” or data from electronic health

records (HER) for genomic discovery
● Regular clinical use of genomics and pharmacogenomics
● Drug discovery based on OMICS data and development
● Individual-level genetic testing to address ethical, legal, and soci-
etal challenges related to such services.
Translational bioinformatics integrates biostatistics, molecular bioinfor-

matics, clinical informatics, and statistical genetics (Chen et al., 2013b).
The field is quickly developing, and many related topics have been
suggested. Among them, pharmacogenomics is a branch of genomics
concerned with genetic differences in drug response. This branch is vital
for future precision medicine design. Translational bioinformatics is a
recent subject that has gained significance in the era of personalised
and precision medicine. Based on curating large amounts of scientific
literature, TBI may identify erroneous research, derive fresh insights
into underlying genetic mechanisms of disease, enhance estimations of
pathogenicity of human genetic variants, and find possible new treatment
targets. However, interpreting the data to establish a clinical diagnosis or
treatment plan is far more difficult than sequencing an exome. Many of
the thousands of discovered variations will need to be examined clinically.
Some Mendelian disorders require just one variation to be found and
examined, such as basic Mendelian disorders. Multivariate analysis will be
required for more complicated disorders (such as cancer, diabetes, and
neurodegenerative diseases). Getting accurate results requires asking the
proper questions about patients and diseases, as well as using the right
computational tools. “Translational genomics” is the use of new findings
from the Human Genome Project to enhance diagnostics, prognostics,
and treatments for complex illnesses.
Biology and information technology have merged to create a new area
of translational study called translational bioinformatics (TBI) (Ritchie
et al., 2020). Aside from basic DNA sequence alterations, epigenomic
data now contains information on methylation and histone modifications
as well as DNA methylation (data above the genome). Through informa-

tion technology, it is possible to acquire and analyse the proteome (the
total amount of proteins in the cell, tissue or organism), transcriptome
(the total amount of mRNA in a cell), and metabolome (complete set of
small molecules, called metabolites, in the cell). Bioinformatics aims to
characterise and quantify molecular groups that contribute to an organ-
ism’s structure, function, or dynamics. This means that every person’s
OMICS profile should be linked with their clinical observations, medical
images, and physiological signals.
3 Bioinformatics Interventions
in Translational Research
The National Institutes of Health (NIH) defines translational research as
having two areas of translation. There are two ways to achieve this: one
is to adapt findings acquired during laboratory and preclinical research
to the creation of clinical trials and human studies. It’s also important
to note that the second area of translation involves research targeted at
boosting the adoption of best practices in the community. In translational
research, the cost-effectiveness of preventive and treatment methods is
equally essential. From the scientist to the user, translational research
shifts the emphasis. User involvement is increasingly important in the
translational research paradigm. They have an impact on the priorities
of academics.
3.1 Translational Biomedicine

Translational biomedical research has recently been a hot topic in the
biomedical research field. Translational research seeks to “translate”
current biological knowledge into methods and instruments for treating
human illness. Beyond that, it’s just medical applied research, with its
own obscure name. Translational research covers medical genetics, cancer,
and cardiology. Currently studied hereditary diseases include microorgan-
isms, plants, animals, and humans. Mendelism established genetics aims
to improve our knowledge of basic, preclinical, clinical, epidemiological,
and healthcare research understandings are improved.
42 C. BARUAH ET AL.
3.2 Translational Clinical Research

Several recent studies have highlighted a 20-year gap between top clin-
ical research knowledge and its use in our health system. Translational
science aims to accelerate the application of scientific discoveries in
clinical settings and foster collaboration between researchers and clini-
cians. Preclinical research, clinical trials, and health technology evaluation,
including Alzheimer’s disease and dementia prevention and treatment,
may all benefit from the findings.
3.3 Translational Stroke Research (TSR)

TSR includes fundamental, translational, and clinical research. Modern
methods for evaluation, prevention, treatment, and repair after stroke
and various types of neurotrauma are being developed. Basic and clinical
scientists and doctors alike may benefit from translational stroke research.
This includes neuroscientists and physicians alike. Preclinical and clinical
neuroprotective effectiveness differences have become more concerning
in translational stroke research.
3.4 Translational Neuromedicine

To deliver new treatments with quantifiable results to patients with
neurological disorders, translational neurology studies all technological
advancements. Conceived to help those at risk of or suffering from
neurological illness convert the vast amount of fundamental neuroscience,
neuropathogenesis, and neuroengineering knowledge into treatments and
quantifiable benefits, bringing together fundamental and clinical neurosci-
entists, Translational Neuroscience aims to improve our understanding of
brain anatomy, function, and illness.
3.5 Translational Oncology

Work in both the laboratory and the clinic to improve oncology patient
care. Clinical trials evaluating new treatment paradigms for cancer are the
outcome of Translational Oncology research. In addition, it includes the
most sophisticated clinical tests of both traditional and novel cancer treat-
ments. Research in Translational Cancer Treatment promotes facilities for
cancer treatment and notable programmes in relevant areas to combined
interdisciplinary and translational cancer control companies. Laboratory

discoveries are translated into new cancer treatments for patients via trans-
lational cancer research (TCR). Because this study often leads to effective
treatments for patients as quickly as possible, it is a boon to society.
Researchers develop instruments for clinical trials with the use of clin-
ical observations made by doctors. Conversely, doctors utilise clinical
observations to guide their efforts.
3.6 Translational Imaging

Clinical and scientific applications of biomedical imaging. These methods
are being used in clinical investigations to uncover chemical imbalances
linked with serious mental disorders and drug addictions. To evaluate
traits in animal models, imaging can detect them, and vice versa. These
studies attempt to “translate” ideas from micro-imaging laboratories to
preclinical settings.
3.7 Discovery Biology

In the areas of cancer and neglected diseases, discovery biology performs
basic and applied drug discovery research. The field of discovery biology
is concerned with basic and practical drug discovery research, especially
in cancer and other neglected diseases. Discovery biology services include
biosafety and ad hoc virus testing (Yin et al., 2021).
3.8 Medical Biotechnology

Medical biotechnology focuses on developing technologies for the
health and pharmaceutical industries. Biotechnology in medicine enables
researchers and doctors to find new drugs and prevention of diseases.
Most medical biotechnologists work in academia or industry. Biotech-
nologists have discovered novel medicines and developed and tested
diagnostic technologies to treat and prevent disease. Depending on
their expertise, medical biotechnologists work in academia or industry.
Academic biotechnologists help medical researchers conduct tests,
whereas industrial biotechnologists create pharmaceutical drugs and
vaccines. Medical biotechnology has created microbial insecticides, insect-
resistant crops, and environmental cleaning techniques.
44 C. BARUAH ET AL.
3.9 Orthopaedic Transition

Translational research in orthopaedics is a fast-expanding area. Cellular
and molecular research must be used properly in the therapeutic envi-
ronment to really enhance people’s health. In addition to bringing
cutting-edge information to the forefront, this project will enable pioneers
of orthopaedic translation to share and mutually improve skills.
3.10 Translational Stem Cell Medicine

To improve the therapeutic use of cellular and molecular biology of stem
cell, Stem Cell Translational Medicine (SCTM) was STEM CELLS Trans-
lational Medicine will assist improve patient outcomes by bridging stem
cell research and speeding translation of new lab findings into clinical
trials. The research of stem cells has developed quickly, yet at an aston-
ishing pace. This chapter aimed to provide a comprehensive list of stem
cell types.
3.11 Translational Proteomics

Translation to decipher complicated disease processes, proteomics uses
multidisciplinary methods. It emphasises fast distribution of new find-
ings. It untangles complicated disease processes utilising multidisciplinary
methods. Proteins are essential components of the physiological metabolic
processes of cells and are essential components of living organisms. Most
human diseases are caused by functional protein interaction dysregulation.
In recent years, advancements in science and technology have enabled the
study of protein interactions inside cells.
3.12 Translational Neuroscience

Developing novel treatments for neurodegenerative, neuropsychiatric,
and developmental disorders is the goal of Translational Neuroscience.
Brain anatomy and function research influence the development of novel
treatments for neurological disorders. It is the process of bringing
new treatments with quantifiable results to neurological illness patients.
Conceived to help those at risk of or suffering from neurological illness
convert the vast amount of fundamental neuroscience, neuropathogen-

esis, and neuroengineering knowledge into treatments and quantifiable
benefits.
3.13 Molecule Therapy

The term “molecular therapy” refers to molecular alterations in cells.
Vaccine development, preclinical target validation, clinical trials, and
safety/efficacy studies are all subjects addressed in the study. Molecular
targeted treatments utilise drugs to target particular molecules on the
surface or within damaging cells. These chemicals help provide signals
to cells to divide or grow. The medicines work by slowing the growth
and spread of cancer cells while sparing healthy cells. Targeted treat-
ments employ several medicines with varying effects. Scientists are trying
customised treatments on both animals and people (clinical trials). But
just a few targeted treatments have FDA clearance. The long-term efficacy
and safety of targeted therapy are unknown.
4 Advances in Translational Bioinformatics

Translational bioinformatics involves in development of storage, analyt-
ical, and interpretative techniques that maximise the conversion of
increasingly large biological and genetic data into proactive, preventative,
predictive, and participatory health. The development of novel techniques
for integrating biological and clinical data, as well as the evolution of
clinical informatics methodology to include biological observations, are
all part of translational bioinformatics research. Butte and Chen popu-
larised the phrase when they published “Finding disease-related genomic
research within an international repository: initial steps in translational
bioinformatics” (Butte & Chen, 2006; Wilson et al., 2022). TBI has
grown in prominence over the last decade, attracting a large profes-
sional community that has published results in high-impact journals and
presented findings at national and international conferences.
Informatics was also emphasised during the annual AMIA Summit
on Translational Bioinformatics, which was hosted in San Francisco.
TBI has been studying how computational tools and techniques may
be utilised to understand, analyse, and manage clinical data since its
inception, advancing the discipline of systems biology in the process. To
assist define this new subject, Altman organised a yearly review session
46 C. BARUAH ET AL.
at the AMIA annual conference (Altman, 2012). Altman emphasised the

scope of translational research in health care, which encompasses illness
treatment, prevention, and monitoring, as well as the discovery and eval-
uation of biomarkers and their application to areas like rare disorders
(Burton & Underwood, 2007) or gene-disease correlations (Caufield
et al., 2022; Denny et al., 2010). Translational bioinformatics, he claims,
combines translational medicine with bioinformatics. Translational bioin-
formatics connects the two disciplines by developing algorithms to analyse
fundamental molecular and cellular data in order to improve therapeutic
outcomes. To enhance patient treatment and our knowledge of biology,
TBI research combines data from molecular (DNA, RNA, proteins, small
molecules, and lipids) and clinical entities (patients) (Altman, 2012).
5 Prospects of Translational Bioinformatics

Precision medicine relies on translational bioinformatics to support
genetic, environmental, and clinical profiles of people, allowing genomic
data to be turned into individualised therapy. The individuals working
in this field tackle the scientific and statistical difficulties presented by
genetic data in an unusual way. Translational bioinformatics includes
clinical genomics, genomic medicine, pharmacogenomics, and genetic
epidemiology. Precision medicine in translational biotechnology has rami-
fications in both clinical and therapeutic areas, such as drug discovery.
Introducing new medicines would require the multimodal cooperation
of clinical personnel, physicians, laboratory staff, biostatisticians, and
bioinformaticians. Clinical genomics assists in the discovery of novel
molecular biomarkers that are verified by clinically relevant genetic testing
(Pagonet al., 2002).
Pharmacogenomics may be concerned with the genomic/clinical
phenotypic connections with pharmacologically active drugs (Rubin et al.,
2005). A few new techniques are being explored in clinical studies.
Drugs for diseases like cancer, AIDS, cardiovascular disease, asthma,
and Alzheimer’s will be created utilising pharmacogenomics. Currently,
pharmacogenomics studies factors that influence a drug’s concentration
reaching its targets. The use of gene expression from cell lines to predict
patient drug response is currently controversial due to cell line variability.
The accuracy of predicting in vivo medication response using a patient’s
baseline gene expression profile varied between 60 and 80%. A mix of
inherited and nongenetic factors affects cancer growth and treatment
resistance. Comparatively to public health and environmental registries,

genetic epidemiology collects genome-based data (Little & Hawken,
2010). It is a process of converting fundamental research into a thera-
peutic environment. The complexity of the human physiology and the
variety of the human population would be a limiting factor for real trans-
lation into clinical practice, but they would provide some inputs to future
medical advances (Hopkins et al., 2021).
5.1 Translational Genomics in Clinical Care

While genetics examines single functioning genes, genomics analyses our
whole DNA, recognising non-coding DNA’s regulatory role and the intri-
cate connections between many genes and the environment. Precision
medicine seeks to promote health and treat illness more accurately by
combining predictive, preventative, personalised, and interactive compo-
nents. Recent advances in fundamental research have revealed new genetic
variations and biomarkers. Over the coming decade, many anticipate
significant advances in genetic testing and genome sequencing.
A genetics medicine service will depend on general practitioners to
assist patients with diagnosis, treatment, and illness prevention. More
doctors may explore adopting genetic testing and genome sequencing in
the future years, with some expecting complete integration into routine
medical treatment within 10 years. It may assist in diagnosis, prognosis,
and therapy. Inhibitors of BRAF and Herceptin® (trastuzumab) are two
examples. PARP drugs are more successful in treating ovarian cancer in
individuals with BRCA gene mutations.
In addition to high-risk DNA variants, thorough genotyping may also
help identify milk and gluten intolerances, as well as mucoviscidosis.
Assembling genetic and HAS data may help uncover low penetrant vari-
ants. MFS is caused by mutations in fibrillin 1 (FBN1). MFS patients
have significant clinical heterogeneity within and between families due to
the disease’s aetiology. TGFBR1, TGFB2, TGFBR2, MYLK1, MYH11,
ACTA2, and SMAD3may assist identify individuals at risk for aortic
aneurysms. Studying these high-risk individuals’ aorta shapes may assist
predict disease progression.
Translational genomics may potentially investigate gene networks of
people with various disorders to learn more about their relationships.
That is why almost half of Down’s syndrome individuals exhibit an over-
protection against heart problems linked with connective tissue. Recent
48 C. BARUAH ET AL.
research indicates FBN1 is increased in Down’s Syndrome (usually down-

regulated in MFS). The creation of genomic networks will help clarify the
connections between various diseases. Understanding linked syndrome
gene networks may lead to targeted gene therapy for illnesses.
Baby with long QT syndrome at Lucile Packard Children’s Hospital
Stanford. In this instance, the baby’s heart stopped many times just after
delivery. Gene mutations may cause Long QT syndrome. Finding the
mutation’s gene is essential to therapy. WGS revealed a previously charac-
terised mutation, as well as variation of a new copy number in the TTN
gene that targeted genotyping alone, would not have detected. It also
took hours or days rather than weeks to get the response.
NGS whole-genome sequencing is vital in the study of complex
diseases like cancer. The fact that drugs work differently in different
patients with same cancer has long been a problem in cancer treatment.
Drugs that target the unique signalling patterns of individual patients
are currently being discovered using large-scale pharmacogenomics and
personal genomics datasets. These include databases for the cancer cell
lines. The NIH’s Cancer Genome Atlas Project analysed the genomic
profiles of over 10,000 people to discover new cancer subtypes. The
variability of drug response is thought to be caused by patients with
specific genomic aberrations. Large-scale datasets can be used to predict
drug combinations, reposition of drugs, and delineate mechanisms of
action. They are becoming increasingly important in drug development.
Precision medicine can thus be tailored to individual patients’ genomic
profiles.
6 OMICs for Drug Repurposing and Discovery

Over the past 60 years, the cost of creating new medicines has increased
significantly, with each new medication costing approximately 80 times
more in 2010 than it did in 1960. The long FDA clearance proce-
dure has also been widely addressed. The time it takes for a lead to be
discovered and approved by the FDA is estimated to be 12 years. Conse-
quently, a growing number of researchers are looking at high-throughput
and computational drug development and repurposing. Recent efforts
have concentrated on utilising omics data, particularly genomics, to iden-
tify novel therapeutic targets and develop new applications for existing
medicines (drug repositioning).
Many new large-scale biological databases, in addition to the Human

Genome Project, will aid researchers in better understanding illness
origins and progressions. Biomolecular structural data can be found in the
RCSB Protein Data Bank, as well as links to other biological resources
including gene and pharmacology databases. Using mass spectrometry,
ProteomicsDB, for example, identifies organ-specific proteins and trans-
lated long intergenic non-coding RNAs in the human proteome derived
from tissues, cell lines, and bodily fluids.
The Human Metabolome Database currently contains approximately
40,000 annotated metabolites entries because of these advancements. It
uses mass spectrometry and NMR spectroscopy to provide experimental
and analytical metabolite concentration data. Databases are believed to aid
in the transformation of clinical practice, specifically in metabolic disor-
ders, such as coronary artery disease and diabetes. In fact, metabolomics
is a rapidly expanding research area that encompasses both endogenous
metabolites as well as chemical and biological substances that interact
with the human body. Compounds from meals, medications, TCM, and
the gut bacterial flora are being fingerprinted by researchers. These will
eventually aid in our understanding of the host–pathogen-environment
connection. These databases aid researchers in improve understanding
the progression of complex diseases. Pattern mining and clustering can
allow for the identification of new biomarkers. Clusters that are partitional
(hard) or hierarchical (tree-like nested structure). These methods may be
sped up by utilising multicore CPUs, GPUs, and FPGAs in parallel.
One way to describe the process of employing an FDA-approved
medicine for a condition other than what it was originally approved for is
“drug repurposing”. Off-label use has mostly been motivated by chance
in the past. Viagra, for instance, was originally designed to treat heart
issues but is now used to treat erectile dysfunction. Early phase clinical
trials are avoided by using a pre-approved medication, saving time and
money.
Association research may lead to the discovery of new pharmaceutical
targets. Sanseau et al. (2012) looked at prior GWAS results and discovered
that 15.6% of them are already pharmaceutical targets (in comparison with
5.7% of the general genome). In 103,638 patients and controls, Okada
et al. discovered 101 overall RA (rheumatoid arthritis) risk loci, 18 of
27 current RA therapeutic target genes, and three approved cancer medi-
cations that might be active against RA. Khatri et al. (2013) identified a
comparable module of 11 genes in eight previous organ rejection datasets.
50 C. BARUAH ET AL.
Scientists discovered two non-immunosuppressive medications that may

be repurposed to regulate these genes in a mouse model. Drug-Gene
Interaction Database (DGI) and PharmGKB are two resources that may
assist in translating genomic research results into effective medications.
The resources for TBI may be found in the Table 1.
Finally, an expanding set of computational and experimental techniques
based on genetic and clinical data enables medication repositioning. An
increasingly wide range of drug repositioning approaches may be used
quickly and efficiently by combining translational bioinformatics, statis-
tical methodologies, chemoinformatics, and experimental procedures.
There are currently efficient techniques for systematic drug repositioning
utilising huge libraries of biologically active molecules. Medicinal chemists
and other translational experts can help reposition drugs.
7 Systems Biomedicine
Systems biology is a new multidisciplinary study that combines biology,
mathematics, computer science, physics, and engineering. Most biolog-
ical systems are too complicated for even the most sophisticated computer
models to capture all system characteristics. A useful mode should be able
to correctly comprehend the system under investigation and give trust-
worthy prediction results. To do this, a certain degree of abstraction may
be needed, focusing on the system behaviours of interest while ignoring
other aspects. Systems biology does not study individual genes or proteins
one at a time, as has been the case for the last 30 years. Rather, it studies
the interactions of all components in a biological system in action. With
the goal of building formal algorithmic models for predicting process
outcomes from component input, systems biomedicine is an emerging
approach to biomedical research. Several important characteristics define
the systems approach:
● Pursuit of quantitative and accurate data

● The datasets’ comprehensiveness and completeness
● Willingness to define, quantify, and alter biological complexity
● Focus on component interconnection and networks
● Obsession with mathematically predicting outcomes.
Table 1 Resources for translational bioinformatics that are open to the public
Name URL Comments
PharmGKB http://www.pharmgkb.org PharmGKB is a curated

resource for physicians and
academics interested in the
effect of genetic diversity
3
on medication response
Phenotype Knowledgebase http://phekb.org Electronic phenotypic
algorithms and their
performance characteristics
may be built, validated, and
shared via an online
collaborative repository
Pharmacogenomic http://www.fda.gov/ Contains a list of
Biomarkers in Drug Labels drugs/scienceresearch/researchareas/pharmacogenetics/ucm083378.htm FDA-approved medicines
that include
pharmacogenomic
information on their labels
Clinical Pharmacogenetics http://www.pharmgkb.org/page/cpic Contain a list of the
Implementation guidelines of CPIC for
Consortium (CPIC) drug-gene interactions
NHGRI Catalog of GWAS http://www.genome.gov/ Curated list of phenotypes
studies and key results of GWAS
studies
Catalog of PheWAS results http://phewascatalog.org Contains the catalogue of
EHR PheWAS results
Drug-Gene Interaction http://dgidb.genome.wustl.edu Data from 13 sources is
database used to provide a search
interface for drug-gene
A REVIEW OF RECENT ADVANCES IN TRANSLATIONAL …
interactions
51
(continued)
52
Table 1 (continued)
Name URL Comments
My Cancer Genome http://www.mycancergenome.org Contains data related to

cancer mutations,
treatments, and relevant
clinical trials
ClinVar http://www.ncbi.nlm.nih.gov/clinvar/ It contains current
C. BARUAH ET AL.
connections between
human variants and
phenotypes, as well as
supporting data
SHARPn http://phenotypeportal.org SHARPn developed a
compendium of computable
phenotypic algorithms
Network studies have been done mostly on cell-based systems like

immunology and cancer, or on homogenous tissues like the heart and
liver. Using new bioinformatics methods, scientists discover microRNA-
gene networks that are important in human inflammatory disorders and
cancer.
7.1 Personalised Genomics

Personalised medicine treatment is crucial for patients to achieve the
best possible outcomes while minimising adverse effects and high direct
medical expenditures. Personalised medicine makes use of genetic and
genomic testing. Whole-genome sequencing (WGS) examines the expres-
sion and interactions of all the genes in the human genome rather than
just one. Rather of providing gene signature profiles based on the expres-
sion levels of individual component genes, genetic testing looks for single
gene mutations or overexpression. Breast cancer genes BRCA-1 and -2,
melanoma gene BRAF, and non-small cell lung cancer gene EGFR are
examples. Breast, colon, and prostate cancer Oncotype DX tests, as well
as the 70-gene test for breast cancer WGS, have grown easier, quicker,
and less expensive since its beginnings. It’s so easy that it might become
a standard test for healthy people in primary care. However, interpreting
WGS results may be difficult.
7.2 Genetic Testing

Genetic testing utilises human DNA, RNA, or proteins to look for gene
variations, chromosomal abnormalities, or proteins linked to illnesses or
disorders. An individual’s risk of getting or passing on a genetic disease
can be assessed using the results of a genetic test. There are already over
1,000 genetic testing available in the United States.
Most medical genetic tests result in changes in medical treatment,
based on data from clinical trials and other medical practice.
● Identify a genetic illness.

● The risk of having a specific genetic disorder.
● Forecast the likelihood of adverse effects or an atypical reaction to a
medication.
● Find a common disease’s elevated risk.
54 C. BARUAH ET AL.
Genomic tests must be accurate and clinically relevant to be useful.

Analytic, clinical, and utility validity are required for genomic tests. In
clinical validation, identifying and quantifying possible causes of biologic
variation is a key aim. Patients gain from tests that are clinically useful.
Using genetic data to drive patient treatment is changing our health care
systems, but there are still obstacles. It examines how people are using
personal genetic information for enhanced diagnoses, tumour profiling,
and genomic risk assessments. Examples of how genomics is being used
in everyday patient care are interwoven with the daily difficulties still
confronting genomics integration into clinical practice, as well as methods
being developed to overcome these hurdles.
7.3 Genomic Testing for Individuals to Inform Health Care

Several companies began offering direct-to-consumer genetic testing in
2008, providing information on genes for both health and leisure. Indi-
viduals can now acquire genetic testing without a doctor’s prescription
thanks to the availability of DTC genetic testing from companies like
23andMe (Mountain View, CA). People received test results as well
as personalised information about their genetic origin, disease risk, and
response to therapy.
DTC genetic testing raises several fascinating ethical, legal, and social
considerations. For years, there was debate on whether or not these exams
should be regulated. In November 2013, the FDA ordered 23andMe
to stop marketing and offering health-related information services. The
FDA designated these tests as medical devices, requiring formal testing
and FDA approval for each test. The FDA accepted 23andMe’s Bloom
syndrome test application in February 2015 (http://www.fda.gov/New
sEvents/Newsroom/PressAnnouncements/UCM435003), and the firm
stated in October 2015 that it would begin giving health information
in the form of 36 gene carrier status. A 23andMe client can download
their raw genetic data and assess the results using information from other
websites, like Geneticgenie, Promethease, Interpretome, and openSNP.
Genetic testing may aid patients, according to a case presented at the
2014 American Neurological Association meeting. Alzheimer’s disease
runs in the family of one of the patients’ mothers. She didn’t know if
she was a carrier or not. She didn’t want to pass that mutation on to her
children, though. Her doctors were able to choose embryos that did not
carry the Alzheimer’s disease gene mutation because to PGD testing. The
patient was never tested, and the number of affected embryos (if any) was
never revealed.
The ability to simultaneously investigate several genes or the entire
genome brings up new possibilities in genomic medicine. Patients are
challenging doctors about the applicability of genetic and genomic
medicine to their own care, since new technologies promise better diag-
noses and treatments. Others believe that incorporating genetics and
genomics into routine clinical practice will be difficult.
7.4 Computational Health Informatics (CHI)

Computational health informatics (CHI) is a relatively recent area of study
in and out of medicine. Information technology (IT) is an interdisci-
plinary that incorporates aspects of biological science as well as medicine.
CHI studies how computers affect health care. Health informatics is the
study of how to forecast a patient’s health by gathering and analysing
data from all areas of healthcare. Patient care is at the heart of health
informatics research (HCO). The amount of medical and healthcare data
available has grown dramatically during the last few years.
The fast development of new technologies has increased the amount of
digital health data in recent years. The digital health data is massive and
complicated in structure for conventional hardware and software. Some of
the reasons why conventional systems fail to handle large datasets include:
● A wide range of organised and unstructured data including medical

records, handwritten doctor notes, MRI, CT, and radiographic films.
● Healthcare informatics has noisy, heterogeneous, complicated,
varied, longitudinal, and big datasets.
● Big data analytics and visualisation challenges.
● The requirement to increase data storage capacity, calculation
capacity, and processing power.
● Improving patient care, data security, sharing, and lowering health-
care costs.
Thus, methods are required to handle and analyse such large, varied,
and complicated information efficiently. Big data analytics, a common
phrase for big and complicated datasets, is critical in handling enor-
mous healthcare data and enhancing patient care. It also has the potential
56 C. BARUAH ET AL.
to save healthcare costs, improve treatments, customise medication, and

assist clinicians in making individualised choices.
Transcriptome and proteome profiling have established how genetic
information is expressed to determine phenotypes. The analysis of the
connection between protein and mRNA levels shows the intricacy of
gene expression regulation during dynamic transitions, particularly during
steady-state and long-term state changes. It is important to note that
the connection between protein levels and coding transcripts is signif-
icantly influenced by mRNA spatial and temporal fluctuation. In this
section, we explain how protein concentrations may buffer mRNA vari-
ation (Liu et al., 2016). Proteomics-based mass spectrometry is a large
and complicated field including numerous mass spectrometers, spectra,
and search results. Quantitation in various scanning modes at various MS
levels adds to the complexity. The most difficult task is quantifying post-
translational modifications (PTM). Many various quantification methods
have been published, some of which may be directly used for PTM quan-
tification (Allmer, 2012). Translation regulates the proteome composition
by converting mRNA coding sequences into polypeptide chains. For
example, translatomics has revolutionised the study of cancer, bacterial
stress response, and biological rhythmicity. The translational design may
increase recombinant protein output by thousands of folds.
8 Related Work
Translational bioinformatics, systems biomedicine, clinical informatics,
statistical genetics, and genomic medicine are all being enticed to play an
increasingly important role in accelerating the translation of genome-scale
studies to hypothesis-driven biological modelling, effective treatment, and
tailored disease management or prevention. Over the last decade, tech-
nological improvements in high-throughput sequencing have resulted in
a growing global capacity for easily creating nucleotide sequences. The
1000 Genomes Project was created in order to compile comprehensive
genetic variation maps of individuals from distinct groups (1000 Genomes
Project Consortium, 2015). For the integration of genetic data with clin-
ical information, data from primary care, hospitals, outcomes, registries,
and social care records should first be gathered using controlled clinical
terminologies such as SNOMED Clinical Terms and the Human Pheno-
type Ontology (Köhler et al., 2017). The Global Alliance for Global
Health (GA4GH) is developing a shared framework of concepts for adop-

tion with the goal of accelerating human health advancements, increasing
efficiency, and lowering costs in order to ensure global interoperability of
medical genetic data (Aronson & Rehm, 2015).
Genomic data enabled the Pan-Cancer Analysis of Whole Genomes
(PCAWG) study, which is an international collaboration aimed at iden-
tifying common patterns of mutation in more than 2,800 cancer whole
genomes from the International Cancer Genome Consortium. PCAWG
proposes to induce genomic, transcriptomic, and epigenomic changes in
50 distinct tumour types and/or subtypes. This research has established
the utility of merging data from several individuals’ genomes, which can
result in the identification of novel targets and disease mechanisms, as
well as improved diagnostic and treatment outcomes for specific patients
(Vamathevan & Birney, 2017).
Large-scale innovative research initiatives, such as the International
Human Cell Atlas Initiative10, which aims to create comprehensive refer-
ence maps of all human cells, as well as devices, applications, wearables,
and implantable technology, will add to this data explosion (Vamath-
evan & Birney, 2017).
8.1 The Ongoing Research Works Are Looking for:

● Developing novel approaches for analysing and merging large-
scale datasets derived from transcriptomic, proteomic, genomic, and
signalling pathways and networks analyses.
● Using machine learning and other modern computational tools to
analyse large-scale biological data sets.
● Identifying and statistically evaluating molecular biomarkers for the
diagnosis, prognosis, and classification of illnesses.
● Making use of cutting-edge bioinformatics technologies such as
Blockchain, Internet of Things, and big data analytics.
● Demonstrates the value of translational bioinformatics approaches in
viroinformatics, drug development, and repurposing.
● Includes translational healthcare and clinical uses of next-generation
sequencing (NGS).
● Demonstrates translational medicine systems and their potential for
healthcare improvement.
58 C. BARUAH ET AL.
● Conducts research on medical image analysis, with a particular

emphasis on CT scans and the detection of novel coronavirus
infections.
8.2 Research Gap

Discovering, reusing, sharing, and analysing data are all dependent on
metadata and data standards in Translational bioinformatics research.
Several recent assessments imply that lack of acceptance of such standards
is often related to difficulties in understanding, accessing, and using them.
(Vamathevan & Birney, 2017).
The informatics literature currently lacks research on “active manage-
ment” of data assets during the lifetime of a research project. Translational
research produces and administers data “in the wild”, meaning researchers
rarely anticipate how the data can be used beyond the original intent.
Leaving data management planning to the last minute diminishes the
value of research data assets and restricts reuse possibilities.
8.3 Future Research Perspective

Patients with genetic disorders account for a sizable proportion of the
world’s population with unique healthcare demands. Recent estimates
place the incidence of chromosomal diseases at 3.8 per 1000, single gene
disorders at 20 per 1000, and multifactorial disorders at 646.4 per 1000.
For example, in India, around 2% of new-borns have single gene or
chromosomal problems, and 3% of couples have children with recurrent
illnesses. Around 30% of major chronic illnesses and over 1,400 common
single gene disorders have genetic variations (Chakrabarty et al., 2016).
The laboratory service should be required to deliver cutting-edge genetic
and genomic testing and analysis. For the reasons stated above, the spec-
trum of variations/mutations in faulty gene(s) is not uniform across the
population, necessitating the use of modern genomic technology. Preci-
sion medicine aims to apply the right dose of the right treatment to
the right patient at the right time. This requires integrating cutting-
edge genetic technologies into healthcare system. Integrative perspectives
on fundamental principles of genomic, proteomic, and computational
biology help researchers as they apply qualitative systems methods to
medical challenges (Tiberti et al., 2022).
9 Conclusion
The biological and healthcare industries are now covered by transla-
tional bioinformatics, which bridges the gap between bioinformatics
and medical informatics. Current research is used to connect biological
data with clinical informatics in translational bioinformatics. It requires
analysing and sequencing an organism’s whole genetic code, from genes
to transcripts. The translational bioinformatics databases help biologists
to learn about disease management and therapeutic development. A
signalling pathway’s operation in a cell can be studied using systems
biology. A comprehensive view of a biological entity is created by
combining genomic, proteomic, and bioinformatics data. Bioinformatics
methods can be used to replicate the appearance of specific human
diseases or healthy states. HGP discoveries are used in translational
genomics to improve the diagnosis, prognosis, and therapy of compli-
cated disorders. These innovations have revolutionised both healthcare
and biomedical research. New tools and methodologies are needed to
turn massive databases into usable knowledge.
References
1000 Genomes Project Consortium, Auton, A., Brooks, L. D., Durbin, R. M.,
Garrison, E. P., Kang, H. M., Korbel, J. O., Marchini, J. L., McCarthy, S.,
McVean, G. A., & Abecasis, G. R. (2015). A global reference for human
genetic variation. Nature, 526(7571), 68–74. https://doi.org/10.1038/nat
ure15393
Allmer, J. (2012). Existing bioinformatics tools for the quantitation of post-
translational modifications. Amino Acids, 42(1), 129–138. https://doi.org/
10.1007/s00726-010-0614-3
Altman, R. B. (2012). Translational bioinformatics: Linking the molecular world
to the clinical world. Clinical Pharmacology & Therapeutics, 91(6), 994–1000.
http://doi.wiley.com/10.1038/clpt.2012.49
Aronson, S. J., & Rehm, H. L. (2015). Building the foundation for genomics in
precision medicine. Nature, 526(7573), 336–342. https://doi.org/10.1038/
nature15816
Buccitelli, C., & Selbach, M. (2020). mRNAs, proteins and the emerging princi-
ples of gene expression control. Nature Reviews. Genetics, 21(10), 630–644.
https://doi.org/10.1038/s41576-020-0258-4
Burton, J. L., & Underwood, J. (2007). Clinical, educational, and epidemio-
logical value of autopsy. Lancet (london, England), 369(9571), 1471–1480.
https://doi.org/10.1016/S0140-6736(07)60376-6
60 C. BARUAH ET AL.
Butte, A. J., & Chen, R. (2006). Finding disease-related genomic experiments

within an international repository: First steps in translational bioinformatics.
AMIA. Annual Symposium Proceedings. AMIA Symposium, 2006, 106–110.
Caufield, J. H., Sigdel, D., Fu, J., Choi, H., Guevara-Gonzalez, V., Wang,
D., & Ping, P. (2022). Cardiovascular informatics: Building a bridge to data
harmony. Cardiovascular Research, 118(3), 732–745. https://doi.org/10.
1093/cvr/cvab067
Chakrabarty, S., Kabekkodu, S. P., Brand, A., & Satyamoorthy, K. (2016).
Perspectives on translational genomics and public health in India. Public
Health Genomics, 19(2), 61–68. https://doi.org/10.1159/000442518
Chen, J., Zhang, D., Yan, W., Yang, D., & Shen, B. (2013). Translational
bioinformatics for diagnostic and prognostic prediction of prostate cancer
in the next-generation sequencing era. BioMed Research International, 2013,
901578. https://doi.org/10.1155/2013/901578
Chen, J., Qian, F., Yan, W., & Shen, B. (2013b). Translational biomedical infor-
matics in the cloud: Present and future. BioMed Research International, 1–8.
http://www.hindawi.com/journals/bmri/2013b/658925/
Denny, J. C., Ritchie, M. D., Basford, M. A., Pulley, J. M., Bastarache, L.,
Brown-Gentry, K., Wang, D., Masys, D. R., Roden, D. M., & Crawford, D.
C. (2010). PheWAS: Demonstrating the feasibility of a phenome-wide scan
to discover gene-disease associations. Bioinformatics (oxford, England), 26(9),
1205–1210. https://doi.org/10.1093/bioinformatics/btq126
Denny, J. C. (2014). Surveying recent themes in translational bioinformatics: Big
data in EHRs, omics for drugs, and personal genomics. IMIA Yearbook, 9(1),
199–205. http://www.ncbi.nlm.nih.gov/pubmed/25123743.
Gochhait, S. et al. (2021). A machine learning solution for bed occupancy
issue for smart healthcare sector. Journal of Automatic Control and Computer
Science (6). Springer, ISSN: 0146–4116.
Hopkins, M. M., Ibanez, F., & Skingle, M. (2021). Supporting the vital role
of boundary-spanning physician researchers in the advancement of medical
innovation. Future Healthcare Journal, 8(2), e210–e217. https://doi.org/
10.7861/fhj.2021-0091
Khatri, P., Roedder, S., Kimura, N., De Vusser, K., Morgan, A. A., Gong, Y.,
Fischbein, M. P., Robbins, R. C., Naesens, M., Butte, A. J., & Sarwal, M.
M. (2013). A common rejection module (CRM) for acute rejection across
multiple organs identifies novel therapeutics for organ transplantation. The
Journal of Experimental Medicine, 210(11), 2205–2221. https://doi.org/10.
1084/jem.20122709
Köhler, S., Vasilevsky, N. A., Engelstad, M., Foster, E., McMurry, J., Aymé,
S., et al. (2017). The human phenotype ontology in 2017. Nucleic Acids
Research, 45(D1), D865–D876.
Little, J., & Hawken, S. (2010). On track? Using the human genome epidemi-
ology roadmap. Public Health Genomics, 13(4), 256–266. https://doi.org/
10.1159/000279627
Liu, Y., Beyer, A., & Aebersold, R. (2016). On the dependency of cellular
protein levels on mRNA abundance. Cell, 165(3), 535–550. https://doi.org/
10.1016/j.cell.2016.03.014
Liu, Z. X., Cai, Y. D., Guo, X. J., Li, A., Li, T. T., Qiu, J. D., Ren, J., Shi, S.
P., Song, J. N., Wang, M. H., Xie, L., Xue, Y., Zhang, Z. D., & Zhao, X. M.
(2015). Yi chuan = Hereditas, 37 (7), 621–634. https://doi.org/10.16288/
j.yczz.15-003
Mohabatkar, H., Rabiei, P., & Alamdaran, M. (2017). New achievements
in bioinformatics prediction of post translational modification of proteins.
Current Topics in Medicinal Chemistry, 17 (21), 2381–2392. https://doi.
org/10.2174/1568026617666170328100908
Pagon, R. A., Tarczy-Hornoch, P., Baskin, P. K., Edwards, J. E., Covington,
M. L., Espeseth, M., Beahler, C., Bird, T. D., Popovich, B., Nesbitt, C.,
Dolan, C., Marymee, K., Hanson, N. B., Neufeld-Kaiser, W., Grohs, G. M.,
Kicklighter, T., Abair, C., Malmin, A., Barclay, M., & Palepu, R. D. (2002).
Genetests-geneclinics: Genetic testing information for a growing audience.
Human Mutation, 19(5), 501–509. https://doi.org/10.1002/humu.10069
Ritchie, M. D., Moore, J. H., & Kim, J. H. (2020). Translational bioinformatics:
Biobanks in the precision medicine era. Pacific Symposium on Biocomputing,
25, 743–747.
Rubin, D. L., Thorn, C. F., Klein, T. E., & Altman, R. B. (2005). A statistical
approach to scanning the biomedical literature for pharmacogenetics knowl-
edge. Journal of the American Medical Informatics Association, 12, 121–129.
https://doi.org/10.1197/jamia.M1640
Sanseau, P., Agarwal, P., Barnes, M. R., Pastinen, T., Richards, J. B., Cardon,
L. R., & Mooser, V. (2012). Use of genome-wide association studies for
drug repositioning. Nature Biotechnology, 30(4), 317–320. https://doi.org/
10.1038/nbt.2151
Vamathevan, J., & Birney, E. (2017). A review of recent advances in transla-
tional bioinformatics: Bridges from biology to medicine. Yearbook of Medical
Informatics, 26, 178–187.
Yin, D., Ling, S., Wang, D., Dai, Y., Jiang, H., Zhou, X., Paludan, S. R.,
Hong, J., & Cai, Y. (2021). Targeting herpes simplex virus with CRISPR-
Cas9 cures herpetic stromal keratitis in mice. Nature Biotechnology, 39(5),
567–577. https://doi.org/10.1038/s41587-020-00781-8
Wilson, A. C., Chiles, J., Ashish, S., Chanda, D., Kumar, P. L., Mobley, J. A.,
Neptune, E. R., Thannickal, V. J., & McDonald, M. N. (2022). Integrated
bioinformatics analysis identifies established and novel TGFβ1-regulated genes
62 C. BARUAH ET AL.
modulated by anti-fibrotic drugs. Scientific Reports, 12(1), 3080. https://doi.

org/10.1038/s41598-022-07151-1
Tiberti, M., Terkelsen, T., Degn, K., Beltrame, L., Cremers, T. C., da Piedade,
I., Di Marco, M., Maiani, E., & Papaleo, E. (2022). MutateX: An automated
pipeline for in silico saturation mutagenesis of protein structures and structural
ensembles. Briefings in bioinformatics, bbac074. Advance online publication.
https://doi.org/10.1093/bib/bbac074
CHAPTER 4
Application of Bioinformatics in Agricultural

Pest Management: An Overview
of the Evolving Technologies
Bhabesh Deka, Azariah Babu, and Uma Dutta
1 Introduction
In all living things of this universe, sustenance is a common threat.
According to Malthus’ renowned essay, food production grows in an
arithmetic ratio while the population expands geometrically. Humans
have understood from the beginning that they must cultivate crops and
expand output to feed an ever-growing population. Since the beginning
of agriculture, men have attempted to raise the uneven curve of food
production, and this curve has tended to rise, barring famines, in the past.
B. Deka (B) · A. Babu

Department of Entomology, North Bengal Regional Research and
Development Centre, Nagrakata, West Bengal, India
U. Dutta
Department of Zoology, Cotton University, Guwahati, Assam, India

https://doi.org/10.1007/978-981-19-6506-7_4
64 B. DEKA ET AL.
Despite the green revolution’s tremendous success and the introduction

of contemporary technology, many people continue to die of hunger, and
a sizable percentage of the world’s population remains hungry. To fight
the ever-increasing threat of pests and diseases, we need to take a step
back and look for fresh solutions (Delmer, 2005). Among all of these
pests, arthropods, which are classified as insects, are the most destructive.
Vector-borne diseases represent the biggest threat to all, even though we
are exposed to a wide range of ailments throughout our lives. Insects
are the primary disease carriers of many such diseases. Many mosquito-
borne diseases continue to kill humans, including malaria, dengue fever,
chikungunya, and filariasis. Numerous epidemics have occurred in trop-
ical and subtropical countries as a result of vector-borne diseases in the
last several years. Pest management science is an area of research that has
evolved through time to address this issue. Pest control in the modern
era constitutes a broad variety of disciplines requiring both basic and
applied sciences to be successful. Integrated Pest Management (IPM) is
the name given to this joint endeavour. As a result of severe bug infes-
tation in certain countries such as the United States of America, from
grains to cash crops, pests feed on a broad variety of goods and materials.
According to the Indian situation, between 2005 and 2007, bud necrosis
virus, American bollworm, Helicoverpa armigera, Spodoptera litura, red
hairy caterpillar, and tobacco caterpillar, among other important preda-
tors, caused significant losses in sunflower, soybean, sugarcane, wheat,
and apple. Because of this, not only does it contribute to the global food
crisis, but it erodes the foundation of our rural economy as well. Several
areas of computational biology are directly or indirectly engaged in pest
control research. Many prominent bioinformatics applications have been
examined and discussed in detail throughout this chapter.
2 An Assessment of the Current

State of Pest Control
Biological and chemical techniques, together with the traditional knowl-
edge that has been accumulated over the years, form the important
components of an integrated pest control strategy. To fight introduced
pests, conventional biological control programmes rely on predators,
illnesses, and parasites native to the pest’s origin. As a consequence of
this approach, a large population is needed to avoid genetic diversity loss
caused by population size bottlenecks. While an introduction is taking
4 APPLICATION OF BIOINFORMATICS IN AGRICULTURAL … 65
place, the effective population size of the originating population may tend
to decrease. A small part of the population is restricted to guarantee any
undesirable illnesses or pest species are eradicated. Due to the small popu-
lation size, the process may take several generations, and it may result in
inbreeding (Murty & Banerjee, 2011). In reality, only a small percentage
of individuals released into a new range contribute to the following gener-
ation, reducing the size of the initial population. Thus, even if the initial
collection is sufficiently enough to prevent severe bottlenecks, genetic
diversity may decrease with time.
In the beginning, chemical pesticides were regarded as a miracle answer
to the growing pest problem by farmers who suffered substantial losses
due to insect-related crop damage. A surge in demand for chemical
pesticides resulted as a consequence of farmers abandoning conventional
pest control techniques by adopting this new strategy. Pesticides used
incorrectly have led to unintended effects in the ecosystem. In recent
years, concerns have been raised about biomagnification, or the build-
up of xenobiotics that are more concentrated in the body than in the
environment (Wiratno et al., 2007).
Risky microorganisms are not only bad for our major basic require-
ments but also damage the economical status of a country (Cavicchioli
et al., 2019). Some of the world’s poorer countries, which depend on
agricultural growth, are bearing the brunt of this economic catastrophe.
Several times attack by insects in India has resulted in similar condi-
tions. States like Andhra Pradesh, Karnataka, Punjab, Uttar Pradesh,
Maharashtra, and Haryana have been particularly hard-hit by recent
year’s drought. There are just a few prevalent pests, including stem
borer, fruit and shoot borer, pod borer, and top shoot borer. It has
been decided to take a few measures to guarantee that our harvest
would survive. Central Plant Protection Stations (CPPSs) and Central
Surveillance Stations (CSSs) were combined in 1992 to become Central
Integrated Pest Management Centres (CIPMCs).
It has been a long time since India has made any significant attempts
to solve these serious issues in tackling the pest to save crops. There are
31 CIPMC in 28 states and one union territory that have been created
to tackle the problem. During the years 2007 and 2008, pest control
incurred a large amount of money. Among the important areas where
additional contributions were considered were the pests monitoring (8.16
million acres), the bio-control chemical field releases (1900 million acres),
and the area coverage (7.00 million acres).
66 B. DEKA ET AL.
In addition to training and monitoring, this huge amount of money has

been utilised to develop innovative ways of combating epidemics. Genetic
(Handler & Beeman, 2003), molecular (Bhattacharyya & Bhattacharyya,
2006), and theoretical methods have lately been proposed and evaluated,
and researchers across the world are working on increasingly sophisticated
and complex ways. Computational biology will likely play a major role in
this field shortly, if not already.
3 Information Technology:
Scope in Biological Sciences
A branch of contemporary science that is multidisciplinary in nature,
bioinformatics, or computational biology, uses information technology
to address biological problems. Mathematical modelling and statistical
techniques were employed for a long time to forecast a wide variety of
biological features, but with the advent of computers, the area saw a
significant transformation. In recent years, methods from the discipline
of bioinformatics have become more popular in all areas of fundamental
biology. Several of the most frequent applications include phylogenetic
analysis, comparative sequencing, structure prediction, and discovery and
validation of drug and pesticide targets (Banerjee et al., 2008; Munjal
et al., 2018). There is a growing demand for data mining to shed light
on previously undiscovered aspects of complicated biological processes
as more data is made available to the general public (Abdel-latief et al.,
2007; Munjal et al., 2018). An additional advantage of this cutting-edge
technology is the ability to create databases with efficient data input and
retrieval methods for data entry (Gochhait et al., 2021).
Database technology has emerged as a key component of information
technology in recent years. Relational databases are extensively used in
a broad variety of areas. Biology databases are more difficult to maintain
than other databases because the data is so complex, and it requires a high
degree of integration to handle and link a range of input files. Primary and
secondary database systems have been developed and are now widely used
in business and academia, as well as in the armed forces.
As a result, academics and researchers have access to hundreds of useful
databases on the Internet, which are accessible for free or for a nominal
fee, databases are becoming more useful in pest management because of
the establishment of relevant and helpful pest information. It’s worth
mentioning that, one such database is Spodobase (Nègre et al., 2006;

Punia et al., 2021), which has data on Spodoptera ESTs.
To produce the ESTs, five unique cDNA libraries derived from three
different species of S. frugiperda were utilised. These contain fat body,
midgut, and haemocytes samples from the S. frugiperda, together with
the Sf9 cell line. Because they are vital for biological processes such as
development, immunity, and insect-plant interactions, these tissues have
been chosen. Current ESTs in the SPODOBASE database total 29,325
and are organised into redundancy-free groups (2294 clusters and 6103
singletons). More ESTs from S. frugiperda or other species may easily be
added to the SPODOBASE as required or when new data becomes avail-
able. To locate information, text searches, pre-formatted queries, query
help, and blast searches may all be utilised. On the other hand, GO-Slim
is utilised to annotate both NCBI and Bombyx mori EST databases using
the same GO-Slim vocabularies.
4Integrated Pest Management

(IPM) Using Information
and Communication Technologies (ICT)
The goal of integrated pest management (IPM) is to keep pest numbers
below the level that causes economic damage. IPM, the most successful
system approach pest control technique, has reaped enormous benefits
from information and communication technology (ICT) applications.
FAO’s IPM Programme Development Officer, Andrew Bartlett, has clas-
sified the many areas of IPM in which ICTs are being used (Xia, 2003). It
can also help researchers analyse data and create scientific papers (Bartlett,
2002). These documents may now be emailed to farmers, extension
workers, researchers, other scientists, and others, or posted on a public
website. Scientists might utilise the Internet to great effect in coun-
tries like India, where Internet kiosks and smartphones are reaching rural
regions. Several features of ICT make it appropriate for this purpose: ICTs
are effective because they are quick, persistent, comprehensive, open,
and interactive. It benefits from Internet forums and websites that allow
users to ask questions. Farmers, government agencies, non-governmental
organisations (NGOs), and other private organisations that advise and
educate farmers to need a higher level of expertise and competence than
68 B. DEKA ET AL.
previous extension systems. ICTs may be utilised to enhance decision-

making and training in IPM (Bartlett, 2002). Identifying pest species,
and their life cycle, scouting, monitoring, establishing ETL, choosing
and deploying control measures, and assessing treatment efficacy may all
be done using ICT. Initiatives like IPM CRSP, Africa IPM Link, and
IPMNet have developed ICT information channels such as discussion
forums, websites, email list servers, online databases, and multimedia CD-
ROMs. A list of available environmentally friendly bio-control methods
and a pest distribution model linked to weather monitoring systems are
included in this decision support system.
5 Entomo-Informatics: A Prelude
to the Concepts in Bioinformatics
Entomo-informatics is a scientific field that is essential in today’s ento-
mological studies. Data collected in sequencing facilities has led to a
new field called bioinformatics. Many biologists are currently unaware
of bioinformatics methodologies, tools, and databases, which may lead
to missed opportunities or misunderstanding of data. Due to pesticide
resistance, entomological research has increased relevance. It represents
different entomological databases, and suitable their URL addresses.
These databases include all proteins, biological reactions, and physiolog-
ical processes. The NCBI (Benson et al., 2013), DDBJ (Tateno et al.,
2002), and EMBL (Calabrese, 2019; Stoesser et al., 2002) are important
shared platforms for storing biological data. The NCBI’s Entrez platform
is a powerful biological search engine that helps us find material. With so
much research going on globally, an International Nucleotide Sequence
Database Consortium (INSDC) was created to minimise data duplication
in the previously stated genome databases (Arita et al., 2021; Cochrane
et al., 2010; Stevens, 2018). NCBI’s sequence databases receive genomic
data from worldwide sequencing efforts and are the backbone of bioinfor-
matics research. There are several nucleotide databases grouped under the
nucleotide database category. The Entrez search engine returned substan-
tial data for the term insecta. Researchers interested in individual genes
will go farther, sequencing them and submitting them to the various
repositories.
DroSpeGe is a search engine for Drosophila genomes. Drosophila
genomes are made accessible for comparative investigations on this plat-
form (Gilbert, 2007; Song et al., 2011). DroSpeGe is a tool for biologists
to compare species differences and similarities, including genes, genome

structure and evolution, and gene function relationships. It also includes
the Blast search tool to find comparable sequences when new assemblies
are added to genome maps. DroSpeGe includes a tool called BioMart for
mining annotations and sequences. FlyBase is a Drosophila-only database
with all the latest genome sequencing news (Marygold et al., 2013).
Drosophila researchers use this database for gene annotations, pheno-
typic data, and expression data. GIFTS (Gene Interactions in the Fly)
are a database of gene interactions in Drosophila pattern development.
DRES has a collection of Drosophila EST sequences. The most researched
creature has pioneered the way for scientists, as a model for geneticists,
developmental and molecular biologists. The medically important pest
and most dangerously considered are mosquitoes, which are vectors for
serious diseases such as yellow fever (A. aegypti), Chagas disease (R.
prolixus ), malaria (A. gambiae), elephantiasis (C. pipiens ), and typhus
(the body louse P. humanus ).
6 Pest Control Decision Support System

It is a set of algorithms that uses a corpus of information to attain
expert-level competency in explaining problems (Banerjee et al., 2008).
There is a need for a knowledge base and an inference engine in every
expert system (ES). This is due to the wide range of applications of
expert system technology to biological issues and it is very difficult to
categorise them. There are several reasons to use an ES system in pest
control monitoring support systems. When disaster assessment and early
warning circumstances arise, ES may be utilised as a decision-making
tool. It has been utilised to solve a wide range of agricultural issues
using knowledge-based expert system technologies. Nilaparvata lugens,
a brown planthopper, have been a major pest in Shanghai’s rice-growing
region since the early 1980s. A variety of weather conditions in tropical
and temperate regions affect this pest’s prevalence and dispersal rates. By
draining nutrients from the phloem, BPH promotes leaf drying and tiller
wilting. The importance of decision aids for evaluating the harm caused
by this insect is essential for a successful battle against this bug and effec-
tive management methods. In this regard, an interactive WebGIS expert
system for early warning of BPH disasters was created, including data on
a variety of variables such as pest density and meteorological conditions.
The system categorises disasters using if–then logic. The server-to-client
70 B. DEKA ET AL.
WebGIS browser offers early warning results and disaster management

methods for BPH pandemic parameters given in the interactive WebGIS.
PREDICT is yet another example of ES’s full potential being fulfilled.
EXSYS and INSIGHT2 + were used to create a portable expert system
for assessing insect damage to red pine stands in Wisconsin. For foresters
who are not familiar with forest pathology or entomology, PREDICT
is a useful tool. There are 28 distinct pathogens, including mammalian,
insect, and disease species in addition to two different types of abiotic
damage, that PREDICT detects. Farmers and agribusiness professionals
may utilise the Greenbug Expert System to manage green bugs and other
cereal pests in winter wheat by using the Greenbug Expert System. On the
Department of Entomology and Plant Pathology’s website, one may find
this expert system at http://entoplp.okstate.edu/greenbug/index.htm.
Students will learn to identify and study common cereal aphids; identify
and study natural enemies; determine treatment thresholds; and choose
a pesticide from a list of available pesticides. An increasing number of
industries are being affected by computers, smartphones, biotechnology
products, the Internet, and so-called “smart” equipment. Private compa-
nies and public and/or private research are bringing these technologies
to the agricultural sector. Pest control as part of integrated agriculture
production is changing recognition to bioinformatics. GMOs, integrated
systems, and precision agricultural equipment that provide site-specific
IPM are three emerging technologies that are anticipated to have an
impact on IPM adoption.
7 Integration of Agricultural
Systems and Pest Biology
There is a dire need for new pest control techniques. Agricultural systems
and pest biology expertise are required to practice sustainable agriculture.
To assist in agricultural production management, models are computer-
based, such as decision support systems (DSS) that integrate data and
human knowledge. For example, Colorado State University is developing
a Pest Management Decision Support System in the United States, and
the work has already begun. To achieve this, a system was developed
in cooperation with all other sectors of the food and fibre manufac-
turing industry. Producers, academics, and policymakers now have a
better understanding of how to monitor and sustain an integrated model
system. Fertility and pesticide inputs may be controlled on farms to main-

tain appropriate levels for groundwater and the ecosystem as a whole by
utilising the many components of this system.
8 Identification of the Target

Protein and Its Structural Analysis
When it comes to the time to increase crop yields and post-harvest quan-
tities, chemical pesticides and biological pesticides are important. As of
right now, pesticides are the most commonly utilised method of pest
management systems. In the end, the target receptor or protein molecule
determines whether a pesticide is effective or not (LópezPazos & Sala-
manca, 2008). The pesticide’s effectiveness depends on its ability to target
a certain insect is the first and foremost factor. Second, the pesticide must
neither damage the crops or indirectly the consumers of such goods,
and third, the target must be sufficiently reliable to eliminate the pest
effectively and accurately. The enzymes chitinase, chitin synthase (Wang
et al., 2019; Yang et al., 2021) are very effective targets for pesticides.
As a result of this, target specificity should be addressed cautiously in
the meantime, it is crucial to understand the structure and behaviour of
the target protein molecule. For example, X-ray crystallography, nuclear
magnetic resonance, and mass spectroscopy are all techniques used to
identify the structure. They are also quite expensive and take a lot of
time. These efforts have been made possible in large part using in-silico
technologies. The simplest approach to get insight into a target struc-
ture is to use existing prediction algorithms. The ab-initio technique
(Deka et al., 2021; Levine, 1991) and homology modelling methods
(LópezPazos & Salamanca, 2008; Haddad et al., 2020) are useful and
practical for comprehending a target molecule. Gibbons (1985) attributes
the former to graph theory and the latter to homology. A common
ancestor must exist for all species since all life began as a single cell. Struc-
turally linked molecules will have both conserved and changing regions in
common with each other. It is considered to be above the twilight zone
if two sequences have more than 25% identity between them.
9 Target Pesticides and Their Virtual Screening

The pharmaceutical and biological sciences have grown more depen-
dent on discovering new drugs to treat patients. Calculative methods are
72 B. DEKA ET AL.
important in this very successful field and it is possible to detect pesticides

using the same technique. A few important technologies have advanced
rapidly in this area. When an HIV protease inhibitor was discovered,
molecular modelling methods gained traction (Wlodawer & Vondrasek,
1998) and are two pillars that support the target-oriented screening
method.
10 Computational Chemistry and Docking

Docking is an essential technique in computational chemistry for
screening new pesticides. Among virtual screening methods, this method
has attracted a lot of interest. For modelling receptor-ligand complex
stability and its three-dimensional structure and is considered a wonderful
tool. It is first necessary to investigate the conformational space of the
ligands that bind to the target molecules, and then the set is scored based
on anticipated binding affinity. There are a lot of different conformations
of the ligand–protein complex that may be generated using this tech-
nique. An algorithm’s docking type is determined by the search technique
by (a) while docking, search for conformational space, (b) before docking,
search for conformational space, or even (c) use incremental docking.
The first set of algorithms optimises the shape and orientation of the
tiny molecule inside the receptor-binding pocket. This linked optimisation
problem is difficult to apply to large datasets because of its complexity.
Monte Carlo, simulated-annealing, or genetic algorithms are frequently
employed to deal with this problem (Lorenzen & Zhang, 2007).
11 Quantitative Structure–activity
Relationship (QSAR)
Over the past four decades, QSAR has been widely utilised in agro-
chemistry, chemistry, toxicology, and pharmaceutical chemistry (Hansch
et al., 2001; Kwon et al., 2019). As a result of rigorous testing and inde-
pendent variable fine-tuning, the output of molecular and atom-based
descriptors, as well as those produced by quantum chemical calcula-
tions and spectroscopy, has risen (Cho, 2005). It is now possible to
screen a large number of substances under the same test conditions
using high-throughput screening techniques. This reduces the risk of
combining findings from different sources. It’s time to go back to the
basics. QSAR techniques are now multidimensional, ranging from 0 to
6 dimensions, depending on the application and validation methodology

(Visco et al., 2002). Because it enables us to evaluate the importance
of structural, geometric, thermodynamic, electrostatic, and other descrip-
tors in connection to the biological activity of a chemical series, this
sophisticated approach provides a technological complement to docking
techniques. Both docking and QSAR are effective techniques for finding
the most promising pesticide candidate chemical (da Silva Mesquita et al.,
2020; Isyaku et al., 2020; Li et al, 2007). New pesticides with fewer
negative effects are increasingly in demand. Through the use of virtual
methods, we have made progress by decreasing the amount of time it
takes to conduct experiments. These cost-effective, time-efficient, and
accurate techniques form the basis of the contemporary approach to
problem-solving point of view.
12 Pest Management: Role

of Mathematical Modelling
Efforts have been made to model pest-affected areas spatiotempo-
rally, as well as implement strategies to remove the danger. A wide
range of advanced optimisation methods and artificial intelligence-based
approaches are being utilised to solve different pest management prob-
lems (He et al., 2019; Karar et al., 2021). As a result of theoretical
biological applications, pesticides have been manufactured more effi-
ciently and more effectively. Artificial neural networks (ANNs) (Banerjee
et al., 2008; Kujawa & Niedbała, 2021) and GA (Genetic Algorithms) are
increasingly used in various optimisation fields. Researchers have shown
that to maximise pesticide crystal protein (PCP) production on a large
scale by using the Bacillus thuringiensis (Hofte & Whiteley, 1989); it
is essential to regulate the concentrations of different processes vari-
ables, in particular, during fermentation. It has also being investigated if
other pest control techniques based on the same bacteria might be used
(Khasdan et al., 2007). A basic quadratic regression equation is used to
connect the output variables to the process input variables in fermentation
optimisation techniques. As a result of biological systems being so compli-
cated, traditional methods are useless. Therefore, data-driven optimisation
techniques are approaches, such as ANNs.
Various researches have shown that artificial intelligence techniques
may be used to achieve different type’s pest control goals. A model of the
worldwide distribution of insect pests was created using self-organising
74 B. DEKA ET AL.
maps (SOM) to detect geographical patterns and species assemblages

(Gevrey et al., 2006). It was determined that each species posed a certain
invasion risk using the SOM weights, allowing researchers to assess a
species’ potential to spread to a certain geographic area. Now, pest control
uses more sophisticated mathematical modelling methods. As a result
of their increased efficiency and accuracy, these methods are seldom
employed in isolation. The Support Vector Machine (SVM) and heuristic
techniques are two of these methods (Doran & Ray, 2014; Meyer et al.,
2003). The quantitative structure-retention relationship (QSRR) was
used to estimate the retention durations (RTs) of 110 different pesticides
or toxicants (Li et al., 2007). There have been many complicated tech-
niques created for assessing the probability of pest establishment following
the introduction of a species into a region (Isman, 2019). Graphic tech-
niques and computer-based decision-making tools like BIOSECURE (a
biosecurity risk management tool for New Zealand’s indigenous ecosys-
tems) (Barker, 2010) were included. No comparison study has ever been
done to further understand the best approach selected. As part of the
investigation, a probabilistic estimate was made of the pest’s likelihood of
developing at each location based on a comparative study. The chosen
method evaluated the likelihood of establishment before the pest was
introduced. The consequence is a false awareness of pest threats before
a real attack. As a consequence, many simulation methods aid in the
creation of pesticide manufacturing, risk assessment for pest assault, and
other jobs that are essential for developing an effective plan during a
genuine attack or prevention.
13 Advanced Strategies
There is a lot of promise for microRNA-related techniques in a variety of
emerging technologies (Kim & Nam, 2006). When it comes to digesting
mRNAs, this amazing molecule is more precise. A pest’s unwanted gene
product will be much simpler to target, and pest population manage-
ment will become significantly more precise in the foreseeable future.
These techniques are used to study and produce pharmaceuticals (Ford,
2006). Apart from that, it may also be used to kill insects. The scientific
community has not yet found this region. Computational biology has the
potential to be very beneficial in miRNA screening (Pla et al., 2018).
In the past, various microRNA screening techniques have been devel-
oped; however, this area still needs a great deal of research and specificity.
However, the application of information technology in pest management

may be expanded. On the other hand, Benfenati et al. (2003) suggested
log p values for 235 pesticides. Whether experimental or virtual, a single
value (log P) should not be relied upon; instead, log P values should be
calculated using many different techniques and the most consistent results
should be used.
14 Key Issues and Major

Challenges and Their Solutions
The advancement of Next Generation Sequencing (NGS) technology
has transformed the field of genomics, allowing for the rapid, cost-
effective, and accurate generation of sequencing data. These technologies
are widely employed in a wide range of sectors, including pest control.
However, because no one expected the rapid expansion of sequencing
technology, computational biology is confronted with significant hurdles
in data storage, administration, security, and analysis (Jayaram & Dhingra,
2010; Kushwaha et al., 2017). Since the invention of DNA sequencing
in 1977, the technology has experienced a significant reduction in
sequencing prices. Unfortunately, computers are not advancing at the
same rate. Lower cost, and hence larger scale, of genome sequencing
generates massive volumes of data, causing serious central processing
unit (CPU) and storage issues. Since the handling of such massive data
sets necessitates computing resources beyond the capabilities of a typical
computer, there are two solutions. The conventional solution is to employ
a computer cluster. This is not cheap, since it necessitates investments
in hardware, software, physical storage space, and energy and cooling
costs for the cluster. Moving calculations to the cloud is a newer, more
cost-effective option. Cloud infrastructures are adaptable and dynamic,
allowing users to adjust assigned resources up and down as needed.
Because the cloud is managed by a third party, there is no need to be
concerned about hardware or related expenditures (Jayaram & Dhingra,
2010; Kushwaha et al., 2017).
Among the most difficult challenges for bioinformaticians are the
search for and use of publically available data, as well as coping with
diverse formatting styles. A tremendous quantity of data is publicly avail-
able, yet using it might be difficult. After finding and downloading the
data, it is critical to analyse it, ensure its quality and applicability, and avoid
losing all of the information in the process. These are time-consuming
76 B. DEKA ET AL.
procedures that frequently result in blunders. Researchers estimate that

around 80% of their time is spent on data preparation and only 20% on
real data analysis. This is due to a lack of defined file formats and uneven
data formatting, which means that each new software generates a different
data format. The idea is to automate these time-consuming and difficult
data-grooming procedures, freeing up researchers’ time to focus on data
analysis. When data is submitted to a “format-free” data analysis platform,
it “loses” the format and becomes a meaningful biological entity, with
all objects of the same sort responding equally, regardless of underlying
formatting variations. Other prevalent concerns include repeatability and
organisational challenges, such as wrongly identified genes or a complete
absence of data annotation. Keeping track of data provenance is critical,
and specifics such as scripts or exact versions of data used must be metic-
ulously documented so that the analysis can be replicated by someone
else or in the future. Because reproducibility is essential for cumulative
science, researchers should pay close attention to such issues (Kushwaha
et al., 2017). Another issue that researchers worry about is losing track
of their own computations, making it extremely difficult to duplicate the
process on another set of data. Keeping track of data provenance is one
approach to assure data repeatability. Noting down all scripts and param-
eters, on the other hand, takes time. As a result, automation would be
a huge benefit here, saving researchers a lot of time and effort. Data
curation and the usage of limited vocabulary are two topics in agricul-
tural bioinformatics that require attention. Editing scientific data is vital
for information transmission; consequently, well curated datasets must be
regularly created through analysis by experts in the area, with the findings
made available to the public. There is a need for collaborative research
groups to exchange restricted vocabularies. The plant ontology (PO) and
gene ontology (GO) consortia’s work would aid in the consistent deploy-
ment of limited vocabulary databases. This will allow for the quick sharing
of knowledge and information about agricultural pest management issues
and techniques. It would be beneficial to link agricultural knowledge
resources.
15 Future Prospects
Although computational biology has made a significant contribution to
pest management via its numerous services, there is still more work to be
done in this area. The number of specialised databases, such as Spodobase
(Nègre et al., 2006), is small in contrast to the hundreds of databases

available for other subjects. Insect spatial–temporal modelling is another
significant field, and while many new techniques are being developed,
many more are still required which will take a lot of learning to track pest
movements over time. However, the integrated decision support system at
Colorado State University is a great example of real and successful use for
computational biology or core information technology. India is still in its
infancy when it comes to computer science and information technology.
Still more work to be done in terms of development and strategy. Several
agricultural research institutes and universities should take the lead in
developing such a comprehensive approach to agricultural development.
Numerous research institutions and universities in India are trying to
adapt to the present technological change to provide the best pest control
help possible. Indian authorities maintain tabs on mosquito species and
the diseases they transmit (Murty et al., 2006). Using sequence features
to group mosquito species has improved Anopheles classification, iden-
tification, and vectorial capability assessments (Banerjee et al., 2009).
According to a new study, agricultural pest microRNA target prediction
has been performed recently (Gualtieri et al., 2020; Jike et al., 2018;
Moné et al., 2018; Singh & Nagaraju, 2008).
16 Conclusion
Future research in pest control should focus more on this important topic.
Numerous significant uses of current multidisciplinary fields linked to
computational biology to pest control are described in this article. The
fact that the study is interdisciplinary lends credence to the idea that
there is still a lot to be done. However, this is slowly changing. Due to
the urgency and importance of the situation, several studies are presently
being performed. Involvement from mathematicians and statisticians is
growing and there is optimism that a global solution could be found
soon if, they work together. These disciplines’ combined efforts may one
day help us find out a solution that does not need us to give up our
food freely, or even a small part cheerfully. For this reason, we must posi-
tion ourselves as the fittest to ensure our survival via the use of technical
instruments in a situation where the survival of the fittest scenario occurs.
78 B. DEKA ET AL.
References
Abdel-latief, M. (2007). A family of chemoreceptors in Tribolium castaneum
(Tenebrionidae: Coleoptera). PLoS ONE, 2(12), e1319.
Arita, M., Karsch-Mizrachi, I., & Cochrane, G. (2021). The international
nucleotide sequence database collaboration. Nucleic Acids Research, 49(D1),
121–124. https://doi.org/10.1093/nar/gkaa967
Banerjee, A. K., Arora, N., & Murty, U. S. N. (2009). Clustering and classifica-
tion of anopheline spacer sequences using self organizing maps. The Internet
Journal of Genomics and Proteomics, 4(1).
Banerjee, A. K., Kiran, K., Murty, U. S. N., & Venkateswarlu, Ch. (2008). Classi-
fication and identification of mosquito species using artificial neural networks.
Computational Biology and Chemistry, 32, 442–447.
Barker, K. (2010). Biosecure citizenship: Politicising symbiotic associations and
the construction of biological threat. Transactions of the Institute of British
Geographers, 35(3), 350–363. http://www.jstor.org/stable/40890992
Bartlett, A. (2002). ICT and IPM, farmers, FAO and field schools: Bringing IPM
to the grass roots in Asia (pp. 8–9).
Benfenati, E., Gini, G., Piclin, N., Roncaglioni, A., & Vari, M. R. (2003).
Predicting log P of pesticides using different software. Chemosphere, 53,
1155–1164.
Benson, D. A., Cavanaugh, M., Clark, K., Karsch-Mizrachi, I., Lipman, D. J.,
Ostell, J., & Sayers, E. W. (2013). Genbank. NAR D36–42.
Bhattacharyya, S., & Bhattacharya, D. K. (2006). Pest control through viral
disease: Mathematical modeling and analysis. Journal of Theoretical Biology,
238, 177–197.
Calabrese, B. (2019). Standards and models for biological data: Common
formats. In S. Ranganathan, M. Gribskov, K. Nakai, & C. Schönbach (Eds.),
Encyclopedia of bioinformatics and computational biology (pp. 130–136).
Academic Press. https://doi.org/10.1016/B978-0-12-809633-8.20418-4.
Cavicchioli, R., Ripple, W. J., Timmis, K. N., et al. (2019). Scientists’ warning to
humanity: Microorganisms and climate change. Nature Reviews Microbiology,
17 , 569–586. https://doi.org/10.1038/s41579-019-0222-5
Cho, S. J. (2005). Hologram Quantitative Structure-Activity Relationship
(HQSAR) study of mutagen X. Bulletin of the Korean Chemical Society, 26(1),
85–90.
Cochrane, G., Karsch-Mizrachi, I., & Nakamura, Y. (2010). The interna-
tional nucleotide sequence database collaboration. NAR. https://doi.org/
10.1093/nar/gkq1150
da Silva Mesquita, R., Kyrylchuk, A., Grafova, I., Kliukovskyi, D., Bezdudnyy,
A., Rozhenko, A., Tadei, W. P., Leskela, M., & Grafov, A. (2020). Synthesis,
molecular docking studies, and larvicidal activity evaluation of new fluori-

nated neonicotinoids against Anopheles darlingi larvae. PLoS ONE, 15(2),
e0227811. https://doi.org/10.1371/journal.pone.0227811
Deka, B., Baruah, C., & Barthakur, M. (2021). In silico tertiary structure predic-
tion and evolutionary analysis of two DNA-binding proteins (DBP-1 and
DBP-2) from Hyposidra talaca nucleopolyhedrovirus (HytaNPV). Biologia,
76, 1075–1086. https://doi.org/10.2478/s11756-020-00665-x
Delmer, D. P. (2005). Agriculture in the developing world: Connecting innova-
tions in plant research to downstream applications. Proceedings of the National
Academy Science, 102(44), 15739–15746.
Doran, G., & Ray, S. (2014). A theoretical and empirical analysis of
support vector machine methods for multiple-instance classification. Machine
Learning, 97 , 79–102. https://doi.org/10.1007/s10994-013-5429-5
Ford, L. P. (2006). Using synthetic miRNA mimics for diverting cell fate: A
possibility of miRNA-based therapeutics? Leukemia Research, 30, 511–513.
Gevrey, M., Worner, S., Kasabov, N., Pitt, J., & Giraudel, J. L. (2006). Esti-
mating risk of events using SOM models: A case study on invasive species
establishment. Ecological Modelling, 197 , 361–372.
Gibbons, A. (1985). Algorithmic graph theory. Cambridge University Press.
http://pi.lib.uchicago.edu/1001/cat/bib/674902
Gilbert, D. G. (2007). DroSpeGe: Rapid access database for new Drosophila
species genomes. NAR, 35, D480-485.
Gochhait, S. et al. (2021). The comparison of forward and backward neural
network model—A study on the prediction of student grade. Journal of
WSEAS Transactions on Systems and Control, 6. ISSN: 1991–8763, 422–429
(Scopus indexed)—Research Funded by UGC, Nepal.
Gualtieri, C., Leonetti, P., & Macovei, A. (2020). Plant miRNA cross-kingdom
transfer targeting parasitic and mutualistic organisms as a tool to advance
modern agriculture. Frontiers in Plant Science, 11, 930. https://doi.org/10.
3389/fpls.2020.00930
Haddad, Y., Adam, V., & Heger, Z. (2020). Ten quick tips for homology
modeling of high-resolution protein 3D structures. PLOS Computational
Biology, 16(4), e1007449. https://doi.org/10.1371/journal.pcbi.1007449
Handler, A. M., & Beeman, R. W. (2003). United States department of
agriculture-agricultural research service: Advances in the molecular genetic
analysis of insects and their application to pest management. Pest Management
Science, 59, 728–735.
Hansch, C., Kurup, A., Garg, R., & Gao, H. (2001). Chem-Bioinformatics
and QSAR: A review of QSAR lacking positive hydrophobic terms. Chemical
Reviews, 101(3), 619–672.
He, Y., Zeng, H., Fan, Y., Ji, S., & Wu, J. (2019). Application of deep learning
in integrated pest management: A real-time system for detection and diagnosis
80 B. DEKA ET AL.
of oilseed rape pests. Mobile Information Systems, 4570808, 1–14. https://doi.

org/10.1155/2019/4570808
Hofte, H., & Whiteley, H. R. (1989). Insecticidal crystal proteins of Bacillus
thuringiensis. Microbiological Reviews, 53, 242–255.
Isman, M. B. (2019). Challenges of pest management in the twenty first century:
New tools and strategies to combat old and new Foes Alike. Frontiers in
Agronomy, 1, 2. https://doi.org/10.3389/fagro.2019.00002
Isyaku, Y., Uzairu, A., Uba, S., Ibrahim, M. T., & Umar, A. B. (2020). QSAR,
molecular docking, and design of novel 4-(N , N-diarylmethyl amines) Furan-
2(5H)-one derivatives as insecticides against Aphis craccivora. Bulletin of the
National Research Centre, 44(44), 1–11. https://doi.org/10.1186/s42269-
020-00297-w
Jayaram, B., & Dhingra, P. (2010). Bioinformatics for a better tomorrow. Indian
Institute of Technology.
Jike, W., Sablok, G., Bertorelle, G., Li, M., & Varotto, C. (2018). In silico
identification and characterization of a diverse subset of conserved microRNAs
in bioenergy crop Arundo donax L. Science and Reports, 8, 16667. https://
doi.org/10.1038/s41598-018-34982-8
Karar, M. E., Alsunaydi, F., Albusaymi, S., & Alotaibi, S. (2021). A new mobile
application of agricultural pests recognition using deep learning in cloud
computing system. Alexandria Engineering Journal, 60(5), 4423–4432.
https://doi.org/10.1016/j.aej.2021.03.009
Khasdan, V., Sapojnik, M., Zaritsky, A., Horowitz, A. R., Boussiba, S., Rippa,
M., Manasherob, R., & Ben-Dov, E. (2007). Larvicidal activities against agri-
cultural pests of transgenic Escherichia coli expressing combinations of four
genes from Bacillus thuringiensis. Archives of Microbiology, 188, 643–653.
Kim, V. N., & Nam, J. W. (2006). Genomics of microRNA. Trends in Genetics,
22(3), 166–173.
Kujawa, S., & Niedbała, G. (2021). Artificial neural networks in agriculture.
Agriculture, 11, 497. https://doi.org/10.3390/agriculture11060497
Kushwaha, U. K. S., Deo, I., Jaiswal, J. P., & Prasad, B. (2017). Role of bioin-
formatics in crop improvement. Global Journal of Science Frontier Research,
17 (1), 1–13.
Kwon, S., Bae, H., Jo, J., & Yoon, S. (2019). Comprehensive ensemble in QSAR
prediction for drug discovery. BMC Bioinformatics, 20(521), 1–12. https://
doi.org/10.1186/s12859-019-3135-4
Levine, I. N. (1991). Quantum chemistry. Englewood Cliffs (pp. 455–544).
ISBN 0-205-12770-3.
Li, X., Luan, F., Si, H., Hu, Z., & Liu, M. (2007). Prediction of retention times
for a large set of pesticides or toxicants based on support vector machine and
the heuristic method. Toxicology Letters, 10(175), 1–3, 136–144.
LópezPazos, S. A., & Cerón Salamanca, J. A. (2008). Mini review and hypoth-
esis: Homology modelling of Spodoptera litura (Lepidoptera: Noctuidae)
amino peptidase N receptor. Revista De La Academia Colombiana De Ciencias
Exactas, 32(123), 139–144.
Lorenzen, S., & Zhang, Y. (2007). Monte Carlo refinement of rigid-body protein
docking structures with backbone displacement and side-chain optimization.
Protein Science, 16, 2716–2725.
Marygold, S. J., Leyland, P. C., Seal, R. L., Goodman, J. L., Thurmond, J.
R., Strelets, V. B., & Wilson, R. J. (2013). FlyBase: Improvements to the
bibliography. NAR, 41(D1), D751–D757.
Meyer, D., Leisch, F., & Hornik, K. (2003). The support vector machine under
test. Neurocomputing, 55(1–2), 169–186.
Moné, Y., Nhim, S., Gimenez, S., Legeai, F., Seninet, I., Parrinello, H., negre,
N., & d’Alencon, E. (2018). Characterization and expression profiling of
microRNAs in response to plant feeding in two host-plant strains of the lepi-
dopteran pest Spodoptera frugiperda. BMC Genomics, 19, 804. https://doi.
org/10.1186/s12864-018-5119-6
Munjal, G., Hanmandlu, M., & Srivastava, S. (2018). Phylogenetics algorithms
and applications. Ambient Communications and Computer Systems: RACCCS-
2018, 904, 187–194. https://doi.org/10.1007/978-981-13-5934-7_17
Murty, U. S. N., & Banerjee, A. K. (2011). Bioinformatics with solutions in
pest management science: An insight into the evolving technologies. In D.
Reddy Vudem, N. R. Poduri, & V. R. Khareedu (Eds.), Pests and pathogens:
Management strategies (pp. 521–542). BS Publications, CRC Press.
Murty, U. S. N., Rao, M. S., Arora, N., & Krishna, A. R. (2006). Database
management system for the control of malaria in Arunachal Pradesh. India
Bio Information, 1(6), 194–196.
Nègre, V., Hôtelier, T., Volkoff, A. N., Gimenez, S., Cousserans, F., Mita, K.,
Sabau, X., Rocher, J., Miguel, L. F., Emmanuelle, D., Audant, P., Sabourault,
C., Bidegainberry, V., Hilliou, F., & Fournier, P. (2006). SPODOBASE: An
EST database for the lepidopteran crop pest Spodoptera. BMC Bioinformatics,
7 , 322.
Pla, A., Zhong, X., & Rayner, S. (2018). miRAW: A deep learning-based
approach to predict microRNA targets by analyzing whole microRNA tran-
scripts. PLOS Computational Biology, 14(7), e1006185. https://doi.org/10.
1371/journal.pcbi.1006185
Punia, A., Chauhan, N. S., Singh, D., kesavan, A. K., Kair, S., & Sohal, S.
K. (2021). Effect of gallic acid on the larvae of Spodoptera litura and its
parasitoid Bracon hebetor. Science and Reports, 11, 531. https://doi.org/10.
1038/s41598-020-80232-1
82 B. DEKA ET AL.
Singh, J., & Nagaraju, J. (2008). In silico prediction and characterization of

microRNAs from red flour beetle (Tribolium castaneum). Insect Molecular
Biology, 17 (4), 427–436.
Song, X., Goicoechea, J. L., Ammiraju, J. S. S., Luo, M., He, R., Lin, J., Lee, S.,
Sisneros, N., Watts, T. A. D., Golser, K. W., Ashley, E., Collura, K., Braidotti,
M., Yu, Y., Matzkin, L. M., McAllister, B. F., Markow, T. A., & Wing, R. A.
(2011). The 19 genomes of drosophila: A BAC library resource for genus-
wide and genome-scale comparative evolutionary research. Genetics, 1,187 (4),
1023–1030. https://doi.org/10.1534/genetics.111.126540
Stevens, H. (2018). Globalizing genomics: The origins of the international
nucleotide sequence database collaboration. Journal of the History of Biology,
51, 657–691. https://doi.org/10.1007/s10739-017-9490-y
Stoesser, G., Baker, W., van den Broek, A., Camon, E., Garcia-Pastor, M., Kanz,
C., et al. (2002). The EMBL nucleotide sequence database. NAR, 30(1),
21–26.
Tateno, Y., Imanishi, T., Miyazaki, S., Fukami-Kobayashi, K., Saitou, N.,
Sugawara, H., et al. (2002). DNA Data Bank of Japan (DDBJ) for genome
scale research in life science. NAR, 30(1), 27–30.
Visco, D. P., Jr., Pophalea, R. S., Rintoulb, M. D., & Faulon, J. L. (2002). Devel-
oping a methodology for an inverse quantitative structure-activity relationship
using the signature molecular descriptor. Journal of Molecular Graphics and
Modelling, 20(6), 429–438.
Wang, Z., Yang, H., Zhou, C., Yang, W. J., Jin, D. C., & Long, G. Y. (2019).
Molecular cloning, expression, and functional analysis of the chitin synthase 1
gene and its two alternative splicing variants in the white-backed planthopper,
Sogatella furcifera (Hemiptera: Delphacidae). Science and Reports, 9, 1087.
https://doi.org/10.1038/s41598-018-37488-5
Wiratno, D. T., Paul, J. B., Ivonne, M. C. M. R., & Albertinka, J. M. (2007).
A case study on Bangka Island, Indonesia on the habits and consequences of
pesticide use in pepper plantations. Environmental Toxicology, 10, 405–414.
Wlodawer, A., & Vondrasek, J. (1998). Inhibitors of Hiv-1 protease: A major
success of structure-assisted drug design. Annual Review of Biophysics and
Biomolecular Structure, 27 , 249–284.
Xia, Y. (2003). The status of ICTs in integrated pest management, (11), 1–2.
Yang, X., Xu, Y., Yin, Q., Zhang, H., Yin, H., Sun, Y., Ma, L., Zhou, D., & Shen,
B. (2021). Physiological characterization of chitin synthase A responsible for
the biosynthesis of cuticle chitin in Culex pipiens pallens (Diptera: Culicidae).
Parasites Vectors, 14, 234. https://doi.org/10.1186/s13071-021-04741-2
CHAPTER 5
Application of Bioinformatics in Health Care

and Medicine
P. Keerthana and Saikat Gochhait
1 Introduction
The field of biology has advanced into data research discipline and enor-
mous wealth of data created is sequenced and available for synthesis of
new information (Bayat, 2002). Thus, there is a need for careful storage,
organization, synthesis and analysis of large data set for discovery of new
knowledge. Hence, information technology and computation techniques
are applied to biology to create a field called bioinformatics.
P. Keerthana (B) · S. Gochhait

Symbiosis Institute of Digital and Telecom Management, Constituent of
Symbiosis International Deemed University, Pune, Maharashtra, India
S. Gochhait

https://doi.org/10.1007/978-981-19-6506-7_5
84 P. KEERTHANA AND S. GOCHHAIT
Basic aims of bioinformatics are storing biological data in a database,

developing tools to analyse data and interpreting meaningful results. The
bioinformatics comprises computational and application bioinformatics.
The computational bioinformatics deals with development of an appli-
cation using computational work like, algorithm development, software
development, database construction and curation for easy retrieval. Bioin-
formatics combines genetics with genomic technologies in discovering
new clinical applications to estimate correlation between gene sequences
and diseases, predicting protein structure from amino acid sequence and
in discovering new drugs to customize treatment for individuals based on
DNA sequence. The essentials of bioinformatics include development of
software tool and algorithm, analysis and interpretation of biological data.
Bioinformatics tools are used for saving, retrieving, analysing and
development of an efficient algorithm enabling measurement of sequence
similarities (Table 1) (Mehmood et al., 2014; Rai et al., 2012). Bioin-
formatic applications include sequence analysis, molecular modelling,
molecular dynamics, etc.
Bioinformatics has input from several areas of biotechnology and
biomedical sciences. There are extensive variety of application of bio infor-
matics in the arena of biotechnology in which the bioinformatic tools are
used to compare the gene pair alignment that aid in identification of the
functions of genes and genomes. It is also used in the study of molec-
ular modelling, a method for analysing the three-dimensional structure of
biological macromolecules, computer aided drug designing with the help
of molecular docking, annotations, etc.
In genomics, bioinformatics is concerned with sequencing and analysis
of genes (Collins et al., 2003). Structural genomics deals with genome
structure identification, determination and characterization. Functional
genomics focuses identification of genes, based on functions. Nutritional
genomics is concerned with nutritional relevance to identified genes.
Proteomics involves interaction proteomics to identify protein–protein
association and expression proteomics for protein quantification.
Bioinformatics has various applications in the field of medicine ranging
between research and molecular medicine, drug development, gene
therapy and preventive medicines. With application of bioinformatics, new
drug discoveries can be personalized to individual’s genetic pattern. In
5 APPLICATION OF BIOINFORMATICS IN HEALTH CARE … 85
Table 1 Bioinformatic tools may be classified as follows:
Sequence Homology and similarity Protein function Structural

analysis analysis analysis
Align, BLAST, FASTA, ENA, CluSTr, Finger- RASMOL,

CENSOR, SSEARCH/GGSEARCH/GLSEARCH PRINTScan, PyMOL,
SeWeR, Inquisitor, Qutemol,
Dna InterProScan, Ascalaph
Block Phobius, Designer,
Aligner, PPSearch, GROMACS,
Clustal RADAR, Pratt, MDynaMix
W2, GLIMMER, TINKER,
CpG Proteax NAMD, Jmol,
Plot, Swiss-PdbViewer,
MAFFT, MaxSprout,
MAUVE, DaliLite,
HMMER, PDBeMotif,
BioEdit, PDBeFold,
Vec Tempura
Screen,
ORF
Finder,
Pepinfo,
SAPS,
Transeq,
etc
gene therapy, it can be used to manipulate the expression of an individual

gene that have been adversely affected. Bioinformatics is also utilized
in preventive medicine in combination with epidemiology. Drug devel-
opment is one of the major applications of bio informatics along with
computational tools, it aids in analysing the disease process and vali-
date new and cost-effective drugs that target the cause of the disease.
In cheminformatics and drug design, bioinformatics enable identification
and structural modification of natural products and designing a drug with
sought after properties.
Molecular genome, an application, helps in conducting DNA
(Deoxyribonucleic Acid) sequencing and is useful in environmental
clean-up, energy production, industrial processing and hazardous waste
clean-up. Bio informatics in evolutionary studies helps to compare and
determine the genome data of different species and characteristics.
Major application of bioinformatics in various sectors includes:
• Biotechnology • Antibiotic resistance • Forensic analysis of

delivery microbes
• Molecular medicine • Evolutionary studies • Stem cell therapy
• Microbial genome • Climate change studies • Plant modelling
• Gene therapy • Alternative energy • Biofuels
source
• Personalized • Bioweapon • Crop improvement
medicine development
• Preventive medicine • Improving nutritional
quality
• Drug discovery • Insect resistance
Limitations of bioinformatics are errors in sequence alignment, algo-

rithm lacking capability to reflect the reality and exhaustive algorithm
with slow rate computation. Hence, comparing results from different
algorithms provide more accurate predictions.
1.1 Major Challenges in Application of Bioinformatics

The main challenges of bioinformatics include abundance volume of raw
data, aggregate information and exponential growth of knowledge created
by the study of genomes and its expressions. The increase in data volume
poses a challenge in management of data, analysis of data and mining of
data acquired that necessitates the need to adapt methods to address this
knowledge base using formal approach like ontologies. The new knowl-
edge thus acquired has to be in computational form for utilization. These
challenges can be dealt with successfully in collaboration between the
experts from various domains like computer scientists and biologists in
dealing with the issues arising and enhance the application of knowledge
developed in various fields (Lyon et al., 2004).
2 Bioinformatics and Secretome Analysis

Secretome is a collection of proteins secreted by the cells of an organism
into the extracellular space and is essential in all the physiological, devel-
opmental and pathological processes. It includes enzymes, growth factors,
cytokines, chemokine’s, hormones, antibodies, interferon’s and plays a

vital role in cell proliferation, cellular immunity, cell differentiation,
cellular communication and morphogenesis (Mukherjee & Mani, 2013).
As profiling of secretomes has become an indispensable area of
research, an in-depth analysis and availability of computational tools help
in systematic examination of secretome data. Thus, the bioinformatics
helps in mapping and analysing protein sequences and aids in creation
and visualization of 3D structure models and gene annotation.
Secretome analysis of unknown sequences is done by, Serial Anal-
ysis of Gene Expression (SAGE), protein microarray (antibody array
and bead based array), signal sequence traps and mass spectrometry
(Liquid chromatography tandem mass spectrometry (LC–MS/MS). It
utilizes gel-based methods which include two-dimensional gel elec-
trophoresis (2DE) and differential gel electrophoresis (DIGE) followed
by mass spectrometry to identify secretory proteins. The gel-independent
methods like iTRAQ (Isobaric tag for relative and absolute quantiza-
tion), SILAC (Stable isotope labelling by amino acids in cell culture) and
ICAT (Isotope-coded affinity tag) are often followed by LC–MS/MS or
SELDITOF MS (Surface Enhanced Laser Desorption Ionization-Time-
of-Flight mass spectrometry) (Jamesdaniel et al., 2009; Mukherjee &
Mani, 2013).
The software’s widely used in 2DE gel image analysis are Decyder,
Progenesis, Dymension, GelScape, ImageMaster 2D, Melanie, Delta2D,
PDQuest, Flicker, ProteomWeaver and LC/MS image analysis are—
CPM, Decyder MS, MsInspect, OpenMS, OBI-Warp,OpenMS, ChAMS,
PETAL, LCMSWARP, SpecArray, TOPP, MapQuant, Msight, etc. 2DE
gel method is used to detect differential proteins associated with disease
and treatment. An alternative to 2DE is the LC/MS proteomic outflow.
Many software’s are available for protein identification and quantitation
like GutenTag, MASCOT, InsPecT, CPFP, Libra, Msquant, etc.
Secretome analysis for known sequences is done by RNA sequencing,
DNA microarray and bioinformatics. Bioinformatics plays an important
role in predicting secretory protein based on signal peptides. The secre-
tory proteins with signal peptides (SP) predicted using computational
analysis are SignalP and SecretomeP. The advantages of computational
algorithm are identification of the secretory protein with SP or without
SP. The disadvantages include absence of experimental validation, false
positives and false negatives are frequently experienced.
Bioinformatic tools for prediction of secreted proteins are based

on: Weight matrices (Predisi, SigCleave, SpScan, TMpred) Sequence
alignment (Signal-BLAST (Basic Local Alignment Search Tool), WOLF
PSORT) and Machine algorithm (MEMSAT, LipoP, ProtComp, PRED-
Lipo, PRED-TAT, NClassG, SecretomeP, SignalP, SPOCTOPUS,
TargetP, TATAFIND, TMHMM). Public repositories provide list of
secretome profiles and links to external databases (e.g.: Interpro, PMAP,
Pfam, Uniport, Swiss-Prot, PROSITE, Entrez Protein, PRIDE, FSD,
OMIM, etc.). Development of bioinformatics tools and improvement in
data bases facilitates experimental analysis of proteins thus beneficial for
therapeutic purposes (Caccia et al., 2013).
3 Mass Spectrometry and Chemical

Crosslinking Reagents
Mass spectrometry (MS) has advanced as an indispensable aspect
in proteomics towards interpreting structural protein complexes
(Matzinger & Mechtler, 2021). Application of mass spectrometry-based
proteomics includes protein interaction networks, clinical proteomics,
global proteomics, single cell proteomics, post-translational modifications,
etc.
The two strategies widely used in protein identification and character-
ization are bottom -up and top-down approach. In bottom-up approach,
the proteolytic digestion of proteins occurs before mass spectrum analysis.
Intact protein is analysed by fragmentation in top-down approach and no
digestion of protein is required and is time efficient (Chen et al., 2020).
Software tools are available for protein identification (MaxQuant,
PEAKS, Mascot, TopPIC, PepHMM, SWPepNovo, EigenMS,
SEQUEST, PECAN, PepNovo PROVALT, InsPecT, DirecTag, DBParser
MassSieve, etc.) and protein quantification (iTracker, Libra, Label-free,
OpenMS, Perseus, MaxQuant, IsobariQ, PVIEW, Skyline, Proteome
Discoverer, XPRESS, PTMselect, etc.) (Petrotchenko & Borchers, 2010;
Tran et al., 2015).
A number of downstream bioinformatics tools and databases are
available (ProLoc-GO, PSEA-Quant, COVAIN, GNET2, IKAP, INGA,
Pathview, KSEA, CRONOS, viper, STRING, SIGNOR, PANTHER,
etc.) for proteomics analysis.
Crosslinking elucidates the structural information of protein. Intra-
molecular or inter-molecular joining of two or more molecules via
covalent bond with the use of crosslinking reagent is known as chemical

crosslinking. Cross linking reagents are selected based on their chem-
ical reactivity, cell membrane permeability, spacer arm length, chemical
specificity, membrane permeability, cleavability and homo-bifunctional or
hetero-bifunctional reactive groups. Cross linking reagents are very crucial
and will affect the specificity and sensitivity of the analysis (Tran et al.,
2015).
The crosslinking along with mass spectrometry is a powerful method
in yielding structural information of protein and protein complexes (Back
et al., 2003; Iacobucci et al., 2020).
There are many crosslinking mass spectrometry (CXMS) data anal-
ysis tools in use like: Xquest, StavroX/MeroX, Xisearch/XIFDR,
ECL/ECL2, Mango, Kojak, pLink/pLink2, Proteome Discov-
erer/XlinkX SIM-XL, etc. Software tools for visualization to examine
protein interaction networks are ProXL, XVis, CLMSVault, XiNET,
Jwalk/MNXL, XlinkAnalyzer, Xwalk, TopoLink, etc. Information
procured by crosslinking mass spectrometry helps in understanding of
various important biological questions. It is also essential in assessing
pathophysiological processes and aid in development of new drug
discoveries. With the recent advances in bioinformatics tools, various
computational issues of crosslinking mass spectrometry are addressed.
4 Bioinformatics and Software Product Line

Development in technology has led to a large repertoire of data being
generated and it in turn necessitates the need of efficient storage, organi-
zation and retrieval of data. Data handling has become more streamlined
with the arrival of bioinformatics. Many of the users of bioinformatics
tools may not have appropriate training on software usage. As there is
demand in different versions of a similar application in various researches,
reuse of software is recommended. Hence, software product line (SPL)
comes into play with development of common platform for a set of prod-
ucts committed to a particular domain. Software Product Line (SPL)
utilizes the idea of “reuse”. It is a group of systems created from a shared
set of basic artefacts. In some cases, such as bioinformatics applications, it
is desirable to use the SPL method to create a collection of linked software
products (Costa et al., 2015).
The main concept of software product line is utilization of typical

asset base in creation of an analogous set of products by core asset
development, product development and management.
Three product line approaches are:
i. Proactive approach, suitable for organizations that can predict their

product line requirement well in advance.
ii. Reactive approach is an extreme programming approach and is used
when the requirements cannot be predicted beforehand.
iii. Extractive approach reuses existing software’s for quick transitions
without much reengineering.
The software product line constitutes multiple software with develop-

ment of common features from common set of core assets (Northrop,
2002). The software assets include software requirements, architec-
ture, design, documentation, prototypes, etc. The software product line
comprises domain engineering concerned with core assets development
and application engineering dealing with final products according to
requirements of customers.
The Software Product line consist of two modules
1. Client Module
2. Core Module
4.1 Client Module

The Client module is mainly constructed using Web Interface. Web Inter-
face is a technique that allows to interact with software or material on a
distant server using web browser. The web server downloads the content
of the web page, allowing us to interact with it using the browser.
Scientific software’s differ from existing one in the following aspects:
i. The method of creation of scientific software is not in traditional

format.
ii. The software typically is developed by researchers and scientific
experts themselves.
iii. Surveying and needs are both impeded since in an initial phase of
study, they may not look obvious or occasionally even unknown.
4.2 Core Module

As the set of readings are ready, there will be two different analyses. If
there is no availability of any genomic information before, it is necessary
to assemble the reads into a genome or transcriptome (or) if genome is
already available, it’s possible to map our reads based on that genome.
Although these two analyses appear similar, they are quite different.
Mapping is the process of comparing each read to the reference genome
and can get one or more sorts between every reading and genome.
The software product line architecture is essential to assure product
features, organize development, control complexity and manage evolu-
tion. The effectiveness of SPL is based on investigation of the common-
alities and variability among products. There is a need for significant level
of commonality. The variations include time dimension and space dimen-
sion. Huge quantities of data are generated by technological advances
and bioinformatic tools help in aggregating, analysing and interpreting
these data’s. Advantages of systematic reuse of product line utilization
are enhancement of quality, improved productivity and decreased cost
and time (Alvis, 2003). Bioinformatics serves as an important applica-
tion domain involving management and analysis of mounting volume of
varied data.
5 Bioinformatics and Protein Kinase

Healthcare system is confronted with many emerging morbid conditions
where gene has an important role in deciding certain diseases (Gochhait
et al., 2021; Rimal et al., 2021). Most of the diseases are linked to
abnormal kinase activity like: cardiovascular diseases, cancer, bronchial
asthma, hypertension diabetes mellitus, etc. Hence, investigation of kinase
activity is indispensable in the field of drug designing and treatment of
various diseases.
Protein kinase uses Adenosine triphosphate (ATP) to phosphorylate
proteins. There are different groups of protein kinase which are further
classified into families. Protein kinase and its substrates have important
function in various cellular processes, and imperfections leads to diseases;
thus, protein kinase plays vital role in drug discovery.
Different categories of protein kinase are: Serine/threonine protein
kinases (adds the phosphate to serine or threonine), Tyrosine specific
protein kinases (uses tyrosine as acceptor of phosphate), Histidine specific

kinase, Aspartyl/glutamyl protein kinase and Tryptophan kinase.
Mechanism for substrate phosphorylation by a kinase starts with
bonding of Adenosine triphosphate to the binding site of the kinase and
followed by binding of the substrate. Phosphoryl transfer occurs, and
substrate is released from the kinase, and this further leads to Adenosine
diphosphate (ADP) release from the active site.
Two major factors determining protein kinase substrate complex are
substrate recruitment and peptide specificity. The computational methods
aimed at predicting substrate of protein kinases are based on the peptide
specificity (Saunders et al., 2008). Prediction tools include NetwoKIN,
KinasePhos, DISPHOS, PredPospho, Scansite, PPSP and NetPhos. The
bioinformatic tools for analysis of posphopeptide data are simPhospho,
Prophossi, PHOSIDA, PhosFox, etc.
An in-depth study of protein kinase using databases is essential to
explore the regulatory mechanisms and to identify association between
kinases and diseases. Application of bioinformatics is mainly utilized in
analysing structure, function, regulatory mechanism, impact of kinases
on diseases and efficient drug designing. A collection of kinase data
bases are developed with the available protein kinase data to understand
kinase pathways, they include: KKB (Kinase Knowledge Base) a Eidogen-
Sertany’s database of kinase chemical structure and biological activity,
KinBase is a gene resource of protein kinase, Kinweb has compilation
of protein kinases encoded in the human gene, KSD (Kinase Sequence
Database) and KPD (Kinase Pathway Database) (Chen et al., 2015).
Bioinformatics thus helps in elucidating information of patient at
the genetic level and combines genetics with genomic technologies in
discovering efficient drug development with increased efficiency.
6 Bioinformatics and miRNA

Bioinformatics has transformed the field of biology; in fact, it has stream-
lined every aspect of biology from data storage to drug discovery.
Like DNA, proteins and other biomolecules, the world of microRNAs
(miRNA) had been integrated with the infinite potential of bioinfor-
matics.
MicroRNAs are non-coding RNA molecule that is small and single
stranded, regulating gene expression at translational or post-translational
level. Thus miRNA act as key regulators of several cell processes such as
cell proliferation, differentiation, metabolism and apoptosis (Lin et al.,

2021). miRNA are capable of binding to their complementary mRNA’s
and bring about transitional repression either by blocking their access
to protein ribosome complex or by degrading the mRNAs. miRNA’s
have now been emerged as a prospective biomarker for the identification
of diseases like diabetes mellitus, cancer, viral infections, neurodegenera-
tive diseases, etc. With an ever expanding application of miRNA’s, more
research are taking place in this field. Along with the research, next gener-
ation sequencing technology has revolutionized this discipline, furthering
the building up of sequence information. Thus bioinformatics became
indispensable in the world of miRNA’s like any other (Huang et al.,
2017).
Bioinformatics apart from its role in storage of enormous amount
of data plays a pivotal role in miRNA nomenclature assignment, func-
tional annotation, prediction of binding sites, miRNA target interaction,
miRNA expression and biomarker information. A number of databases
and algorithms have been developed towards this (Moore et al., 2015;
Rasheed, 2017). These databases include,
● miRBase: creates a repository to assign stable names to newly found

miRNA’s in a consistent and stable manner.
● miRDB: serves as a repository for miRNA functional annotations
and target predictions.
● miRWalk: clearly predicts the binding sites of miRNA associated with
genes, pathways, cell lines, pathological conditions, etc.
● miRTarBase: presents a comprehensive picture of miRNA target
interaction, to enable graphical visualization of interaction also
contains many other features. It employs word cloud for this purpose
and data source includes GEO, TCGA and CUP-seq.
● miRCancer: is a database designed to provide a comprehensive
information on the expression of miRNA’s in human cancer.
● doRiNA, SomamiR, etc.
● Early Detection Research Network (EDRN): Unlike other databases
which stores sequence based information, EDRN is exclusively
submitted to multiple types of biomarkers.
7 Clinical Bioinformatics and Cancer

Cancer is an abnormal proliferation of cells in an uncontrolled manner
due to mutation, chromosomal rearrangement and error in molecular
machinery. Bioinformatics has a definitive role in various areas of cancer
research like single-cell analysis, cellular imaging, genome sequencing,
biomarker discovery and analysis of transcriptome. Continuous research
in the field of oncology identifies cancer biomarkers, which are a vital
method in research and enable quick diagnosis, therapy and prognosis.
Bioinformatics in research is applied in diagnosing cancer, categorizing
aetiologies, identifying various treatment methods and in assessment
of prognosis. It also helps in understanding mechanism of initiation,
progression and in identification of target. Bioinformatics facilitates
complete analysis by connecting geographically available data of cancer
patients. The data thus collected and organized are available in databases
that allow data access and retrieval.
The international nucleotide sequence database collaboration
(INSDC) covers DNA Data Bank of Japan (DDBJ), European Molec-
ular Biology Laboratory (EMBL) and GenBank at National Centre for
Biotechnology Information (NCBI). These repositories exchange data
among themselves and are updated with the worldwide coverage and
allow full access to all records. Multitude of studies on genome, transcrip-
tome and proteome provides opportunity for investigators to analyse and
integrate available data in cancer research. Many projects are developed to
integrate the data types collected and combined from different sources.
National Cancer Institute (NCI) has launched Cancer Genome Anatomy
Project (CGAP) in cancer genetics, and Human Cancer Genome Project
(HCGP) has developed millions of Expressed Sequence Tags (EST).
These ESTs are integrated into International Database Cancer Gene
Expression, which serves as a human cancer index (Stransky & Galante,
2010).
The NCI has also created a network related to cancer known as cancer
biomedical informatics grid (caBIG) and facilitates in easy sharing of
information to connect scientists and practitioners. It also builds tools
to accumulate, analyse, integrate and disseminate information related to
cancer and treatment. The data for cancer are available from International
Cancer Genome Consortium (ICGC) in coordinating cancer genome
studies and is concerned with the genomic changes in cancer globally, The
Cancer Genome Atlas (TCGA) (Tomczak et al., 2015), National Cancer
Institute (NCI), National Human Genome Research Institute (NHGRI)

and Catalogue of Somatic Mutations in Cancer (COSMIC) (Hinkson
et al., 2017).
The available data can be used in bio simulation, early detection
through identification of cancer biomarkers, risk analysis, identifying
druggable targets and personalized medicine. Bioinformatics tools help
in identifying mutations, molecular markers in different stages of cancer,
profiles used in classification, diagnosis and predict clinical outcomes of
cancer. The application of clinical bioinformatics in cancer is vital in
creation of biobanks and incorporating them with clinical information for
improved therapeutic outcome.
8 COVID-19 Implications
and Sustainable Development Goals
The United Nation has developed an agenda for accomplishing sustain-
able development to provide peace and prosperity in the world. A global
partnership strategy to attain sustainable development, seventeen goals
were prioritized. These sustainable development goals (SDGS) aimed at
removing poverty, reducing inequality, improvement in education and
increased economic growth along with handling climate change and
preserving forest and oceans.
The wide spread infection across the world due to corona virus disease
has caused a huge adverse effect on the humankind. The global pandemic
has caused healthcare crisis and also impacted global economy, thus
hampering the achievement of sustainable development goals by 2030.
The disruption caused by COVID-19 pandemic has led to delay in attain-
ment of agenda set by the SDGS by reversing the progress earlier attained
on hunger, poverty, education, health care, etc (Leal Filho et al., 2020;
Min & Perucci, 2020; Ray et al., 2021).
The primary goal is to end poverty, but COVID-19 has led to
increase in global poverty rate. There is an exacerbation of world hunger
due to the global pandemic with worsening child malnutrition. The
goal of ensuring healthy lives has reverse in progress with short life
expectancy and disruption in healthcare system, leading to increase in
morbidity and mortality during the pandemic. Provision of quality educa-
tion is hampered by the COVID-19 pandemic situation, and the closure
of educational institution has devastating effect on learning and well-
being of students (Gochhait et al., 2021; Rimal et al., 2021). Gender
equality is adversely affected with increased violence against women. In

the pandemic scenario, people worldwide have to live without access to
safe drinking water and proper hygienic service, as these are essential in
containing the spread of COVID-19 infection.
Providing affordable and clean energy is essential to combat climate
changes, which can be achieved by accelerating actions on modern renew-
able energy (Singh & Mishra, 2021). The decent work for all and
sustainable economic growth is highly impacted by the pandemic that
has led to the loss of jobs and is predicted for a substantial increase
in unemployment of youth. The goal of building sustainable industrial-
ization and fostering innovation has plunged as a result of COVID-19
crisis. Reducing inequality within and among countries is one of the goals
of sustainable development that has a reverse progress and has in fact
increased existing inequalities. The pandemic has worsened the quality of
life and forced people to live in poor conditions and slowed down the
goal of sustainable cities and communities. There is an urgent need to
combat climate change, and its impact, the sustainability of oceans, seas
and marine resources are under severe threat and need to be conserved as
this serves as an important aspect of sustainable future.
Degradation of ecosystem has a deleterious effect on human well-
being, and hence, it is important to promote terrestrial ecosystem and
halt biodiversity loss. The global pandemic has exposed that there is a
need for promoting peace and ensure justice for sustainable development.
Trade tension, technological snag and dearth in financial resources have
caused an unexpected blow to the global system and have retarded the
global partnership for sustainable development.
There is an urgent need to invest in data, with innovation in acquiring
data, processing, dissemination and modelling, which is a critical tool for
containing the pandemic and help in risk assessment and prevention. The
data on social and economic impact help to plan for recovery and remedial
actions (Cannataro & Harrison, 2021). The unprecedented crisis and the
severity of the consequence of pandemic is challenge for the government
and communities.
References
Alvis, M. E. (2003). An introduction to software product line development,
26–37.
Back, J. W., de Jong, L., Muijsers, A. O., & de Koster, C. G. (2003). Chemical
cross-linking and mass spectrometry for protein structural modeling. Journal
of Molecular Biology, 331(2), 303–313.
Bayat, A. (2002). Science, medicine, and the future: Bioinformatics. BMJ
(Clinical Research Ed.), 324(7344), 1018–1022.
Caccia, D., Dugo, M., Callari, M., & Bongarzone, I. (2013). Bioinformatics
tools for secretome analysis. Biochimica Et Biophysica Acta, 1834(11), 2442–
2453.
Cannataro, M., & Harrison, A. (2021). Bioinformatics helping to mitigate the
impact of COVID-19—Editorial. Briefings in Bioinformatics, 22(2), 613–615.
Chen, C., Hou, J., Tanner, J. J., & Cheng, J. (2020). Bioinformatics methods
for mass spectrometry-based proteomics data analysis. International Journal
of Molecular Sciences, 21(8), 2873.
Chen, Q., Luo, H., Zhang, C., & Chen, Y. P. (2015). Bioinformatics in protein
kinases regulatory network and drug discovery. Mathematical Biosciences, 262,
147–156.
Collins, F. S., Green, E. D., Guttmacher, A. E., & Guyer, M. S. (2003). A vision
for the future of genomics research. Nature, 422(6934), 835–847.
Costa, G. C., Braga, R., David, J. M., & Campos, F. (2015). A scientific software
product line for the bioinformatics domain. Journal of Biomedical Informatics,
56, 239–264.
Gochhait, S., Butt, S., De-La-Hoz-Franco, E., Shaheen, Q., Luis, D. M., Piñeres-
Espitia, G., & Mercado-Polo, D. (2021a). A machine learning solution for
bed occupancy issue for smart healthcare sector. Journal of Automatic Control
and Computer Science, 55(6), 546–556. ISSN: 0146-4116.
Hinkson, I. V., Davidsen, T. M., Klemm, J. D., Kerlavage, A. R., Kibbe, W.
A., & Chandramouliswaran, I. (2017). A comprehensive infrastructure for big
data in cancer research: Accelerating cancer research and precision medicine.
Frontiers in Cell and Developmental Biology, 5, 83.
Huang, J., Borchert, G. M., Dou, D., Huan, J., Lan, W., Tan, M., & Wu, B.
(2017). Bioinformatics in microRNA research. In Methods in molecular biology
(Vol. 1617). Humana Press.
Iacobucci, C., Gotze, M., & Sinz, A. (2020). Cross-linking/mass spectrometry
to get a closer view on protein interaction networks. Current Opinion in
Biotechnology, 63, 48–53.
Jamesdaniel, S., Salvi, R., & Coling, D. (2009). Auditory proteomics: Methods,
accomplishments and challenges. Brain Research, 1277 , 24–36.
Leal Filho, W., Brandli, L. L., Lange Salvia, A., Rayman-Bacchus, L., & Platje,
J. (2020). COVID-19 and the UN sustainable development goals: Threat to
solidarity or an opportunity? Sustainability, 12(13), 5343.
Lin, J., Zeng, J., Liu, S., Shen, X., Jiang, N., Wu, Y.S., Li, H., Wang, L., &
Wu, J.-M. (2021). DMAG, a novel countermeasure for the treatment of
thrombocytopenia. Research Square.
Lyon, J., Giuse, N. B., Williams, A., Koonce, T., & Walden, R. (2004). A
model for training the new bioinformationist. Journal of the Medical Library
Association, 92(2), 188–195.
Matzinger, M., & Mechtler, K. (2021). Cleavable cross-linkers and mass
spectrometry for the ultimate task of profiling protein–protein interaction
networks in vivo. Journal of Proteome Research, 20(1), 78–93.
Mehmood, M. A., Sehar, U., & Ahmad, N. (2014). Use of bioinformatics tools
in different spheres of life sciences. Journal of Data Mining in Genomics and
Proteomics, 5(2), 1–13.
Min, Y., & Perucci, F. (2020). Impact of COVID-19 on SDG progress
(UN/DESA Policy Brief 18), 1–5.
Moore, A. C., Winkjer, J. S., & Tseng, T. T. (2015). Bioinformatics resources
for DNA discovery. Biomarker Insights, 10(S4), 53–58.
Mukherjee, P., & Mani, S. (2013). Methodologies to decipher the cell secretome.
Biochimica Et Biophysica Acta, 1834(11), 2226–2232.
Northrop, L. M. (2002). SEI’s software product line tenets. IEEE, 19, 32–40.
Petrotchenko, E. V., & Borchers, C. H. (2010). Crosslinking combined with
mass spectrometry for structural proteomics. Mass Spectrometry Reviews,
29(6), 862–876.
Rai, A., Bhati, J., & Lal, S. B. (2012). Software tools and resources for
bioinformatics research (Vol. 1). New India Publishing Agency.
Rasheed, Z. (2017). Bioinformatics approach: A powerful tool for microRNA
research. International Journal of Health Sciences, 11(3), 1–3.
Ray, M., Sable, M. N., Sarkar, S., & Hallur, V. (2021). Essential interpretations
of bioinformatics in COVID-19 pandemic. Meta Gene, 27 , 100844.
Rimal, Y., Gochhait, S., & Bisht, A. (2021b). Data interpretation and visual-
ization of COVID-19 cases using R programming. Informatics in Medicine
Unlocked, 26(6), 100705. ISSN: 0146-4116.
Saunders, N. F., Brinkworth, R. I., Huber, T., Kemp, B. E., & Kobe, B.
(2008). Predikin and PredikinDB: A computational framework for the predic-
tion of protein kinase peptide specificity and an associated database of
phosphorylation sites. BMC Bioinformatics, 9, 245.
Singh, V., & Mishra, V. (2021). Environmental impacts of coronavirus disease
2019 (COVID-19). Bioresource Technology Reports, 15, 100744.
Stransky, B., & Galante, P. (2010). Application of bioinformatics in cancer
research. In W. Cho (Ed.), An omics perspective on cancer research. Springer.
Tomczak, K., Czerwińska, P., & Wiznerowicz, M. (2015). The Cancer
Genome Atlas (TCGA): An immeasurable source of knowledge. Contemporary
Oncology, 19(1A), A68–A77.
Tran, B. Q., Goodlett, D. R., & Goo, Y. A. (2015). Advances in protein complex
analysis by chemical cross-linking coupled with mass spectrometry (CXMS)
and bioinformatics. Biochimica et Biophysica Acta, 1864(1), 123–129.
CHAPTER 6
Information Retrieval in Bioinformatics:

State of the Art and Challenges
Sunita, Sunny Sharma, Vijay Rana, and Vivek Kumar
1 Introduction
The retrieval of biomedical literature is getting increasingly complex,
necessitating the development of improved information retrieval systems.
Unstructured resources, such as text documents, are scoured by Infor-
mation Retrieval (IR) tools in massive data repositories, which are held
on systems. Information depiction, storage, and groups are all phases of
IR (Nadkarni, 2002), in which one of the most complex tasks in IR is
formative which materials are appropriate to the users requirements and
which are not. Users cannot perfectly propose search string in an exact
Sunita (B) · V. Kumar

Department of Computer Science, Arni University, Kangra, India
S. Sharma
Department of Computer Science, University of Jammu, Jammu, India
V. Rana
Department of Computer Science, GNA University, Punjab, India

https://doi.org/10.1007/978-981-19-6506-7_6
102 SUNITA ET AL.
mode to get particular part of data from enormous data reserves under
the current regime. Search results from basic information systems are of
reduced worth. We plan another advancement to filtering searches to
improve imply the user’s data require grouping to work on the results of
IR by utilizing distinctive inquiry extension strategies and produce a linear
arrangement between them, where the linear arrangement was straightly
between two development results all at once in our proposed framework
for this section where the arrangements were linearly between two devel-
opment results at time in our proposed system for this chapter. Query
expansions, for example, discover synonyms and reweight original phrases
to broaden the search query. They deliver substantially more targeted,
specific explore outcomes than standard queries.
The rest of this chapter is facilitated as go after: “related work"
segment gives a diagram of associated effort State of the Art|| Section
discusses the terminologies used in information retrieval. Open Problems
and Challenges|| Section outlines the existing problems and future scope
thereof. Conclusion|| segment is ending, and it also handles on possibility
of future work.
2 Literature
Because of the rapid growth of biological data, good IR systems are
required offers on particular and meaningful responses to complex
queries. One of the key challenges in information retrieval societies is
query extension. Researchers have developed a variety of strategies for
query expansion. Some methods stress the use of unstructured data
(text documents) to determine expansion words, while others emphasize
the use of structured data to determine expansion terms (Ontologies).
Perez Aguera et al. (2010) compare various methods for query expansion
in unstructured documents. The expansion term: Tanimoto, Dice, and
Cosine coefficients to consider co-occurrence of terms in distinct papers.
They also use Kullback Liebler Divergence to look at the allotment of
development expressions in the peak position documents and the entire
collected documents.
Described a work in Buscher et al. (2012) about how to choose
words that are more pertinent to the query topic from feedback docu-
ments based on the placements of keywords in feedback documents. To
solve this challenge in a coherent, probabilistic manner, they developed a
position to importance model (PRM).
6 INFORMATION RETRIEVAL IN BIOINFORMATICS: STATE … 103
The conclusive of their experimentation on two major web-

informational collections uncovers that the recommended PRM is very
effective and hearty, results best in class pertinence models in both reports
and passage-based input. Alipanah et al. (2011) introduced the neces-
sary Expansion Terms (BET) and New Expansion Terms (NET) as new
weighting techniques for ontology-driven query expansion index.
Rivas et al. (2014) developed query-expansion preprocessing strategies
for fetching documents in different domains of biomedical literature from
the Cystic Fibrosis corpus of MEDLINE documents. They performed
tests to demonstrate the varying outcomes and benefits of utilizing
stemming and stop-words in document and query segmentation.
Query expansion has been used in biological information retrieval
research to improve retrieval performance. Abdou and Savoy (2008) used
the SMART retrieval system to test the effectiveness of query expansion
on MEDLINE resources. Literature of biomedical information retrieval,
Xu et al. (2006) observed query-expansion approaches concerning with
local analysis as well as worldwide examination and ontology. Matos
et al. (2010) carry out an archive recovery and prioritization instrument
that dependent on utilizes idea-arranged inquiry extension to discover
reports that are connected to each other. For biological query expan-
sion, Rivas et al. (2014) seem terms query-specific, corpus-specific, and
phrases for language-specific. These investigations showed that by taking
domain-specific factors into account, query expansion is able to improve
efficiency of biomedical IR. According to these investigations, Dang et al.
(2013) used a weighted-dependence model to assign weights to candidate
concepts in order to increase retrieval effectiveness. Xu et al. (2006) used
a blend of significance models to weight query development phrases for
clinical search to discover patient cohorts.
3 State of the Art

3.1 Information Retrieval
Within a big document collection, information retrieval (Krallinger et al.,
2017) is worried about distinguishing a subset of records whose content
is usually applicable to a user’s demand. The purpose of information
retrieval given an immense data set of records and a particular data neces-
sity—normally addressed as a question by the client—is to find the reports
104 SUNITA ET AL.
in the data set that meet the data need. Normally, the task must be
completed correctly and quickly.
Index structures and Boolean queries: A Boolean inquiry (Dadheech
et al., 2018; Du et al., 2020) is a basic and frequent approach to commu-
nicate an information requirement. A term (e.g., OLE1) or a Boolean
term combination is provided by the user (e.g., OLE1 and lipid). The
biomedical literature databases PubMed, as well as many other text search
engines, use this query paradigm.
The vector model and similarity queries (Bordawekar & Shmueli,
2017), is based on give details in this section, is a widely used of Boolean
query. The viewed documents are as (algebraic) vectors over terms in this
configuration, as formally specify below. A search query, q, can contain a
large number of terms and even an entire text. It is also represented as a
vector and is observed as a body of content rather than just a set of search
terms. The retrieval effort is reduced to looking for document vectors
that are the most comparable to the search-query vector in the database.
The several of documents’ similarity measures have been developed and
applied.
3.2 Text Categorization

Text categorization is a task that information retrieval systems frequently
tackle (Tellez et al., 2018). This process is classifying text with category—
tags from a collection of predefined categories. Categorization can be
divided into two techniques. One option is knowledge engineering, in
which the user by hand a set of rules to train expert knowledge about
document categorization. The knowledge acquisition bottleneck is the
key disadvantage of this strategy. A knowledge engineer must interview a
domain expert and manually define rules. Any changes to the categories
necessitate additional participation from the knowledge engineer.
3.3 IR in Biomedical-Informatics
The experimental methods based (Hersh, 2020) that allow for the inves-
tigation of genes and proteins from a whole genome are the initial step
toward molecular comprehension of complex biological processes. While
experimentation are designed and conceded, the ability to observe in the
background of existing information and before hypotheses is critical for
both informed planning and interpretation of outcomes (Rimal et al.,
2021). This kind of information is frequently found in the books. Individ-

uals have had to follow during the literature, article by article, and gene
by gene, in order to find it in the past.
4 Open Problems and Challenges

4.1 Abundant Amount of Information
As previously stated, researchers’ consistent efforts have resulted in a
massive growth in publications in the biological sciences. This volume of
scientific literature necessitates more effort on the part of researchers, who
are often tasked with staying current on all information relating to their
chosen research topics. This endeavor is primarily driven by two factors:
the ongoing rise in scientific outcomes and the absence of correspon-
dence inside life science strengths. Creating fitting procedures, strategies,
and apparatuses to help scientists in crafted by naturally separating the
valuable material from the Web (especially from text sources) has turned
into a significant worry in this situation.
The study subject of bioinformatics literature retrieval and mining is
organically diverse, making the process of finding open problems and
difficulties much more difficult. However, some specific difficulties, in our
opinion, demand more attention from academics than others, resulting
in considerable advances. Let us simply mention a few of them, without
pretending to be exhaustive: I encoding/preprocessing strategies; (ii) the
inherent difficulty of study of retrieval and mining challenges; (iii) princi-
ples and the need for more standardization; and (iv) judgment of existing
tools.
4.2 Encoding/Preprocessing Techniques

The preprocessing approaches (Young et al., 2017) can be classified
roughly along the subsequent dimensions: Natural-Language-Processing
(NLP) (Jang et al., 2021), lexical techniques, and semantic approaches
are the three types of techniques. Currently, NLP cannot ensure that
viable solutions will be developed that can account for the almost limit-
less number of changes in how relevant information is “deployed” in text
documents. Due to its enormous potential, this field may, nonetheless,
become the key relevance in the near future. Lexical approaches that focus
on finding important phrases that can characterize documents are usually
106 SUNITA ET AL.
easier to execute, regardless of whether they are framed in a frequency-

based perspective or not. Actuality, they should only use as an initial stage,
as merely lexical preprocessing, e.g., TFIDF (Dey et al., 2017) does not
appear to be suited for ordinary literature retrieval and mining challenges.
Semantic approaches are somewhere in the center between NLP and
lexical techniques. When using semantic approaches, a common schema is
to addition lexical information with supplementary knowledge, which can
be collecting in a multiplicity of ways. Here are a not many models: I any
text report can be planned to a current taxonomy/ontology, determined
to recognize important ideas and joining them to the actual archive, to
work with additional handling; (ii) explicit term disambiguation strategies
(e.g., inactive semantic ordering, synset examination, or NER investiga-
tion) can be used, to improving the importance of candidate terms, to
be used in further processing; Latent Semantic Indexing (LSI). Blynova
(2019) is a technique that uses singular value decomposition to compute
document and term similarities using a “soft” term matching criterion.
Term and document similarity can be better assessed by the cosine of their
vector expressions when phrases and documents are given as vectors of
statistically independent components. With the introduction of synsets has
grown increasingly prominent. In WordNet (Wang et al., 2020), English
words are organized into synsets, which contain all synonyms connected
to a single topic.
4.3 Fundamental of Literature Retrieval and Mining Problems

Aside from the problems associated with determining the “correct”
encoding/preprocessing approach to use, some activities are intrinsically
tough. Consider a generic open discovery process, framed in the LBD
area that necessitates the selection of the hypothesis to be explored.
Regardless of whether the connected work is coordinated by fitting
heuristics pointed toward narrowing the scope of reasonable specula-
tions, the fleeting intricacy of an open revelation measure stays high,
necessitating the use of specific AI approaches and algorithms.
4.4 Principles and Necessities for Further Standardization

The field of life sciences is rapidly evolving. The purpose, a broad
concurrence among scientists on how to describe biological concepts
is becoming increasingly important. On the one hand, settling names,
contractions, and abbreviations can be trying because of the way that the
equivalent (or comparable) names, truncations, and abbreviations may be
utilized to allude to different might be used to refer to various things.
On the other hand, determining where a composite name begins and
stops in a text might be tricky. These issues, in our opinion, are solely
due to an absence of normal terminology and programming appara-
tuses. Luckily, the Unified Medical Language System (UMLS)—which
unites various well-being and biomedical vocabularies and guidelines
determined to empower interoperability between PC systems—has been
a beneficial endeavor aimed at encouraging uniformity. The UMLS has
three tools: a metathesaurus (which contains terms and codes from
a variety of languages), a semantic network (which allows users to
move between related categories and their relationships), and a specialist
database (equipped with language processing tools). However, due to
a lack of standardization, various other issues remain unsolved. Quite
possibly the most troublesome hardships, as we would like to think, is the
necessity for consequently consolidating writing and organic information
bases. Bioinformaticians’ obligations are particular from those of data set
keepers. Standard apparatuses fit for removing messages and connections
from the writing, just as helping data set custodians in finding pertinent
material for comment, would extensively add to making the issue less
serious or even missing for this situation. Other difficult issues have a
close connection to the organization of scientific publishing
4.5 Measurement of Existing Tools

At present, scientists are trying to figure out how the proposed strategies
might enhance access to existing materials. It has been possible to use new
methodologies to explore and mine the literature through competitions.
A recent study (Dogan et al., 2017; Drost & Paszkowski, 2017) did a
critical examination of text-mining approaches in molecular biology and
demonstrated the potential of this approach (BioCreAtIvE). Academics
from around the world participate in this competition every two years to
compare approaches for
i. detecting physiologically significant entities and their relationships

to existing database entries.
ii. associations between entities and facts. By encouraging researchers
to advance in their specialized areas of interest, further efforts in this
regard could encourage the sharing of important knowledge and
skills.
108 SUNITA ET AL.
5 Conclusion
Bioinformatics text analysis aims to improve access to unstructured knowl-
edge by easing searches, supplying auto-generated summaries, connecting
publications with organized assets, displaying content more visually, and
assisting analysts in the discovery of novel hypotheses. Research in bioin-
formatics text mining has developed over the past few years, from archive
recovery to relationship extraction. There are now a number of tools that
can be used to integrate literature analysis across a range of life science
disciplines, and these tools are being developed at an increasing pace.
As part of this study, we briefly discuss literature retrieval and mining in
bioinformatics, text mining, and writing recovery.
We referenced certain issues worth extra investigation in the second
piece of the paper, fully intent on creating bioinformatics writing recovery
and mining techniques and frameworks. To summarize, the scientific
community is effectively occupied with resolving many issues identified
with writing recovery and mining, and a few arrangements have been
introduced and executed. They will be that as it may, be generally point-
less until mainstream researchers move toward all-inclusive guidelines for
how existing information is published and disseminated with researchers,
with an exacting focus on structure of scientific publications.
References
Abdou, S., & Savoy, J. (2008). Searching in Medline: Query expansion and
manual indexing evaluation. Information Processing & Management, 44(2),
781–789.
Alipanah, N., Parveen, P., Khan, L., & Thuraisingham, B. (2011, July).
Ontology-driven query expansion using map/reduce framework to facilitate
federated queries. In 2011 IEEE International Conference on Web Services
(pp. 712–713). IEEE.
Blynova, N. (2019). Latent semantic indexing (LSI) and its impact on copy-
writing. Communications and Communicative Technologies, (19), 4–12.
Bordawekar, R., & Shmueli, O. (2017, May). Using word embedding to enable
semantic queries in relational databases. In Proceedings of the 1st Workshop on
Data Management for End-to-End Machine Learning (pp. 1–4).
Buscher, G., Dengel, A., Biedert, R., & Elst, L. V. (2012). Attentive documents:
Eye tracking as implicit feedback for information retrieval and beyond. ACM
Transactions on Interactive Intelligent Systems (TiiS), 1(2), 1–30.
Dadheech, P., Goyal, D., Srivastava, S., & Choudhary, C. M. (2018). An efficient
approach for big data processing using spatial Boolean queries. Journal of
Statistics and Management Systems, 21(4), 583–591.
Dang, V., Bendersky, M., & Croft, W. B. (2013, March). Two-stage learning
to rank for information retrieval. In European Conference on Information
Retrieval (pp. 423–434). Springer.
Dey, A., Jenamani, M., & Thakkar, J. J. (2017, December). Lexical TF-IDF:
An n-gram feature space for cross-domain classification of sentiment reviews.
In International Conference on Pattern Recognition and Machine Intelligence
(pp. 380–386). Springer.
Dogan, R. I., Chatr-aryamontri, A., Kim, S., Wei, C. H., Peng, Y., Comeau,
D. C., & Lu, Z. (2017, August). BioCreative VI precision medicine track:
Creating a training corpus for mining protein–protein interactions affected by
mutations. In BioNLP 2017 (pp. 171–175).
Drost, H. G., & Paszkowski, J. (2017). Biomartr: Genomic data retrieval with
R. Bioinformatics, 33(8), 1216–1217.
Du, L., Li, K., Liu, Q., Wu, Z., & Zhang, S. (2020). Dynamic multi-
client searchable symmetric encryption with support for Boolean queries.
Information Sciences, 506, 234–257.
Hersh, W. (2020). Information retrieval: A biomedical and health perspective.
Health Informatics. https://doi.org/10.1007/978-3-030-47686-1
Jang, H., Jeong, Y., & Yoon, B. (2021). TechWord: Development of a
technology lexical database for structuring textual technology information
based on natural language processing. Expert Systems with Applications, 164,
114042.
Krallinger, M., Rabal, O., Lourenco, A., Oyarzabal, J., & Valencia, A. (2017).
Information retrieval and text mining technologies for chemistry. Chemical
Reviews, 117 (12), 7673–7761.
Matos, S., Arrais, J. P., Maia-Rodrigues, J., & Oliveira, J. L. (2010).
Concept-based query expansion for retrieving gene related publications from
MEDLINE. BMC Bioinformatics, 11(1), 1–9.
Nadkarni, P. M. (2002). An introduction to information retrieval: Applications
in genomics. The Pharmacogenomics Journal, 2(2), 96–102.
Pérez-Agüera, J. R., Arroyo, J., Greenberg, J., Iglesias, J. P., & Fresno, V.
(2010, April). Using BM25F for semantic search. In Proceedings of the 3rd
International Semantic Search Workshop (pp. 1–8).
Rimal, Y., Gochhait, S., & Bisht, A. (2021). Data interpretation and visualization
of COVID-19 cases using R programming. Informatics in Medicine Unlocked,
26 (6), 100705. Elsevier, ISSN: 0146-4116.
Rivas, A. R., Iglesias, E. L., & Borrajo, L. (2014). Study of query expansion
techniques and their application in the biomedical information retrieval. The
Scientific World Journal, (1), 1–10.
110 SUNITA ET AL.
Tellez, E. S., Moctezuma, D., Miranda-Jiménez, S., & Graff, M. (2018). An

automated text categorization framework based on hyperparameter optimiza-
tion. Knowledge-Based Systems, 149, 110–123.
Wang, Y., Wang, M., & Fujita, H. (2020). Word sense disambiguation: A
comprehensive knowledge exploitation framework. Knowledge-Based Systems,
190, 105030.
Xu, X., Zhu, W., Zhang, X., Hu, X., & Song, I. Y. (2006, October). A
comparison of local analysis, global analysis and ontology-based query expan-
sion strategies for bio-medical literature search. In 2006 IEEE International
Conference on Systems, Man and Cybernetics (Vol. 4, pp. 3441–3446). IEEE.
Young, N. E., Anderson, R. S., Chignell, S. M., Vorster, A. G., Lawrence, R., &
Evangelista, P. H. (2017). A survival guide to Landsat preprocessing. Ecology,
98(4), 920–932.
CHAPTER 7
Hybrid Support Vector Machine with Grey

Wolf Optimization for Classifying
Multivariate Data
M. Revathi and D. Ramyachitra
1 Introduction
The categorization process consists of the most important aspects
regarding the datamining idea. The classification mechanism has been
discovered to occur frequently in everyday life. For example, in a railway
station, tickets are distributed and classified based on the class required,
in a hospital, patients are classified based on the nature of their disease
and their risk factors (low, medium, and high), in a school, teachers
classify students’ performance based on the grade received (first class,
second class, third class, and fail), and in mobile technologies, (Orriol-
sPuig et al., 2009) the basic goal of multivariate data classification is
M. Revathi (B)
Department of Biotechnology, Bharathiar University, Coimbatore, India
D. Ramyachitra
Department of Computer Science, Bharathiar University, Coimbatore, India

https://doi.org/10.1007/978-981-19-6506-7_7
112 M. REVATHI AND D. RAMYACHITRA
to compute a decision boundary, also known as a class or plane sepa-

ration boundary that divides the input data space into one or more
classes. The nature of the problem domain determines whether a deci-
sion boundary or a class dividing border is a simple straight line, a
sophisticated linear or non-linear representational form. Originally, the
data used on a day-to-day basis was simple in nature and was grouped
together based on the data space’s specific qualities. As a result, obtaining
a judgement boundary class is easier when the data groups are non-
overlapping (Beckett Claire et al., 2017). New models for healthcare
services and other datasets are being incorporated in present medical
practises and industries to minimise the impact of various ailments and
rising healthcare expenses. In the creation of computing environments
in smart hospitals, classification and other data mining methods play a
key role. Classification algorithms, for example, are quite effective in
the classification of patient activity. In the past, computers were used to
construct knowledge-based decision support systems that used domain
knowledge from medical specialists and manually transferred that knowl-
edge into rules or computer algorithms. Furthermore, clinical judgments
were frequently made largely on the doctor’s intuition and experience
rather than the database’s knowledge-rich data (Kalimuthu, Sivanantham,
2021). This practice resulted in unfavourable biases and errors. Exces-
sive medical costs are also a possibility, which may have an impact on
the quality of service supplied to patients. A medical diagnostic might
offer incorrect results in a variety of reasons, including the fault of the
doctor or hospital staff. This procedure takes time and is highly reliant
on the subjective opinions of medical experts. The present paper intro-
duces a grey wolf optimization technique according to support vector
machines (SVM) to gather knowledge automatically from examples or
raw data to address this challenge (Liu et al., 2021). Explain the general
classifier model in Fig. 1 to perform data classification. The design of a
data classifier model, as shown in Fig. 1, begins with the collection of raw
data for any real-world situation, followed by the division of the obtained
data into training and testing datasets (Ros, German et al., 2016). When
the categories of classes under consideration overlap, the primary job is to
find the discriminant function’s optimal point, which reduces the number
of misclassifications in the given data while also minimising the chance
of misclassification of the unavailable data. The primary difficulties that
underpin data classification are generalisation and approximation, based
on this aspect (Kalimuthu, Sivanantham et al., 2021).
7 HYBRID SUPPORT VECTOR MACHINE WITH GREY WOLF … 113
Start
Collect available data from real time

solutions and repositories
Perform training the designed classifier
Test the classifier model developed
Perform evaluation of the classifier modal

developed
Stop
Fig. 1 Steps in a classifier model to perform data classification
To update parameters and learn the classifier structure, the classifier

model uses a percentage of training data from the original datasets and
prior knowledge of the issue area during the training phase. The trained
classifier is evaluated in terms of the percentage of test data by offering a
classification judgement for the considered input pattern during the test
phase. The comparison architecture of existing and hybrid SVM-GWO
algorithms is shown in Fig. 2. To test the suggested technique, you’ll need
a standard multivariate data set. An acceptable UCI data set containing
heart disease diagnosis called “Cleveland, vowel, glass, shuttle, and yeast
data” is used to evaluate the algorithms under consideration. Only a few
entries in the shuttle-2 vs 5 data sets had missing values: yeast-0-3-5-
9 vs 7–8, vowel0, cleveland-0 vs 4, and glass-0-1-4-6 vs 2.During data
cleaning, all such records were eliminated. As a result, the total number
of records is 3365, 1805, 545, 297, and 195.
Multivariate datasets
Normalized data
Separate 70% data for training and 30% data for testing
EXISTING PROPOSED SYSTEM
Random forest, SVM,

adaboost, decision tree Separate 70% training
and 30% testing dataset
70% training data
Local optimal parameter

30% testing data using SVM
Global Optimization
Using GWO
Compare the classification performance of existing and hybrid SVM with

GWO such as Accuracy, sensitivity, specificity, and time period
Fig. 2 Comparison architecture of existing and SVM–GWO
When the input data is normalised, almost all classification methods

improve classification accuracy. The data was normalised after loading
the CAD data by dividing each attribute value by the maximum value
of that particular characteristic in this suggested evaluation method. This
approach converts the data to values between 0 and 1, which are suitable
for almost any classification method. The reminder for this paper is organ-
ised as follows. Section 2 discusses several data classification methods and
investigations. Section 3 describes the suggested hybrid support vector
machine with grey wolf optimization (Pham et al., 2018). In Section 4,
the proposed HSVMGWO and existing support vector machine, random
forest, ada boost, and decision tree experimental results are compared.
The proposed chapter’s final remarks and future scope are found in
Section 5.
2 Literature Review
This present section elaborates a review of previous studies for data classi-
fication using support vector machine-based classifiers (Emamgholizadeh
et al., 2021). Galatenko et al. (2014) gave a formalised definition
related to problem of choosing specification as well as building a
genomic classifier for medical test systems that relies on mathematical
machine learning techniques rather than biological or medical expertise
(Gochhait et al., 2021). Latha (2014) presented a support vector machine
(SVM) for Radial Basis Function technique with automatic analysis of
Magnetic Resonance Image (MRI) (RBF). Karamizadeh Sasan et al.
(2014) published a review on support vector machine (SVM), a pattern
recognition and data categorization algorithm. Using the information
supplied by the support vector machine (SVM), Gürbüz and Kilic (2014)
built a general-purpose, rapid, and adaptive automatic disease detection
system, which enhanced the success rate and reduced the decision-making
time.
Chen Zhi et al. (2016) developed a support vector machine classifier
(GA–SVM) based on a genetic algorithm (GA) for lymph disease diag-
nosis. In the first stage, GA is used to cut the 18 features in the lymph
diseases dataset down to six. A support vector machine with several kernel
functions, such as linear, quadratic, and Gaussian, was used as a clas-
sifier in the second stage. In the field of medical imaging, Lee et al.
(2015) combined support vector machines (SVM) with Active Learning
(AL) into rise prediction based on irregular classes. Jian Xiao and Sheng
Hanmin (2015) proposed a Fuzzy Support Machine (FSVM) as the irreg-
ular problem class (dubbed FSVMCIP), which may be thought of as a
modified FSVM with manifold regularisation for two classes, and there
are two costs associated with misclassification.
QinanJia et al. (2014) proposed a method based on support vector

machines to predict the exact recurrence time (SVM). In the medical
domain, prognosis prediction as a regression issue is frequently used to
anticipate event duration time, such as the duration time of a disease
recurrence. The proposed strategy is compared to four different prog-
nostic algorithms using the Wisconsin Breast Cancer Dataset. Jan Petrich
et al. (2017) studied the effect of different CT scanners on the accu-
racy of high-resolution CT (HRCT) scans in categorising regional disease
patterns in patients with diffuse lung disease using multicenter data.
Bayesian and support vector machine (SVM) classifiers were utilised.
Chang Yongjun et al. (2012) proposed a hierarchical support vector
machine that distinguishes diffuse interstitial lung disease by training a
binary classifier at each node in a hierarchy and allowing each classifier to
use a class-specific quasi-optimal feature set to improve time and accuracy
in computer-aided quantification.
Qiu Wu and Zhang proposed using discriminative information
contained in a whitening transformation to convert LDA to PCA,
followed by a support vector machine (SVM) technique for LDA. The
kernel algorithm for the SVM solution to LDA was constructed by using
the kernel technique to indirectly translate the data to kernel space.
Gudadhe et al. (2010) introduced an assistance vector machine (SVM)
and artificial neural network-based decision support system for heart
disease classification (ANN). To construct a decision support system
for the detection of heart illness, researchers used a three- layer Multi-
layer Perceptron Neural Network (MLPNN). Rokach (2010) examined
existing ensemble techniques and can be used as a tutorial for practi-
tioners who want to create ensemble-based systems. The goal of ensemble
approach is to combine multiple models to create a prediction model.
Based on an SVM and feature selection, Akay (2009) proposed a breast
cancer screening method. In the studies, several training-test partitions of
the Wisconsin Breast Cancer Dataset (WBCD) were used, which is exten-
sively used among researchers that use machine learning algorithms for
breast cancer diagnosis. To tackle the LD diagnosis problem, Wu et al.
(2008) used two well-known artificial intelligence techniques: the arti-
ficial neural network (ANN) and the support vector machine (SVM).
For enhancing the efficiency of support vector machines, Zhan and Shen
(2005) presented a four-step training technique (SVM).
Leng Yan et al. (2016) proposed a new method for improving the
speed and accuracy of SVMs with unlabelled data: one method is to
build SVMs with grid points, which can be expected to speed up SVMs
in the test phase; another method is to build SVMs with unlabelled
data, which has been shown to improve SVM accuracy when there
is very little labelled data. Kim et al. (2003) suggested utilising the
SVM ensemble with bagging (bootstrap aggregation) or boosting to
improve the genuine SVM’s restricted classification performance. To clas-
sify heterogeneous medical data, Kumar and Arasu (2015) employed
Modified Particle Swarm Optimization and Adaptive Fuzzy K-Modes
(MPSO-AFKM). Mishra and colleagues. Hybrid filter-wrapper techniques
for high-performance classification models were developed by (2015).
Nguyen et al. (2015a, 2015b) used wavelet transformation (WT)
and interval type-2 fuzzy logic system (IT2FLS) for automated medical
data classification. The IT2FLS was taught through a hybrid learning
method, resulting in improved performance and reduced computational
burden. Nguyen et al. (2015a, 2015b) proposed a modified Analytic
Hierarchy Process (AHP)-based gene selection for microarray classifica-
tion, with AHP-selected genes being used for cancer classification using
the fuzzy standard additive model (FSAM). In order to handle the
high-dimensional, low-sample nature of microarray data, the number of
fuzzy rules is minimised using a genetic algorithm (GA).In order to
reduce computational burden, Nguyen et al. (2015a, 2015b) proposed
a fuzzy standard additive model (SAM) with genetic algorithm (GSAM)
for healthcare data classification, high-dimensional datasets discriminative
features derived by applying wavelet transformation.
Purwar and Singh (2015) tested a hybrid prediction mode with missing
value imputation on three medical data sets (HPM-MI). Santhanam and
Ephzibah proposed using a genetic algorithm and fuzzy logic to diag-
nose heart illness automatically. To increase the performance of the fuzzy
inference system used to generate a classification model for the GA
selected feature, the fuzzy gaussian membership function and the centroid
technique were applied.
Pourpanah et al. (2019) created a fuzzy ARTMAP (FAM) and Classifi-
cation And Regression Tree (CART) hybrid for medical data classification.
The presented model provides consistent learning, predictions in the
form of a decision tree, and the extraction of valuable explanatory rules
as a decision support tool. Sindhiya and Gunasundari investigated how
the genetic algorithm (GA) and other heuristic methods could be used
to choose characteristics for illness identification in large dimensional
datasets. Stathopoulos and Kalamboukis (2015) demonstrated how to use
latent semantic analysis (LSA) to categorise large medical datasets without

using the SVD solution of the feature matrix.
3 System Design
We describe work done to design and assess ways to handle missing values,
attribute noise, and imbalanced class distribution in datasets to predict in
this research. In this section, we’ll go over a quick overview of the hybrid
approach for hybrid support vector machine with grey wolf optimization
(HSVMGWO) in knowledge discovery. The purpose of this stage is to
select the most appropriate categorization method for a particular dataset.
Because no generalisations about the best classification strategy can be
made, including this phase has necessitated empirical testing of each and
every prediction and analysis for a given dataset. Because the dataset under
study is limited, our suggested approach uses unsupervised learning to
find the best hybrid classification methods.
3.1 Support Vector Machine (SVM) Classifier:

SVM classifiers are based upon the concept of decision planes, which
specify decision boundaries unambiguously. A decision plane does sepa-
rate a bunch of objects with differing class memberships. Figure 3 depicts
a linear classifier with a decision boundary line that distinguishes between
different object types (red and blue objects separately). A certain collec-
tion of items is located on the right side of the decision plane, whereas
another set of objects belonging to a different class is located on the left
side. Several classification tasks are not straightforward, necessitating the
use of more complicated structures in order to achieve optimal separa-
tion and accurately categorise new objects related to the test cases under
consideration (Vapnik, 2000).
The data points will be categorised by allocating the system to closer
of the two parallel planes that are pushed apart as far as possible using
SVM. Standard SVMs, on the other hand, solve quadratic or linear
programmes that take a long time to solve. As a result, PSVM classi-
fiers are being examined, which are meant to do classification with low
computational time using a simple proximity hyperplanes solution algo-
rithm (Samanta & Nataraj, 2009). The SVM idea is explained based on
the sample of a two-class dataset with p points in a q-dimensional real
space, S = {xi , yi }, I = 1, . . . , p, xi ∈ Rq . In xi , the q-dimensional vector
Fig. 3 Linear SVM

classifier with a decision
plane
category yi ∈ {−1, 1}, refer to a certain class. A + and A are allocated

to the[ equivalent datasets
] belonging to classes yi = 1 and yi = −1. The
A = x1 , x2 , . . . , x p , T and D = diag(yi ). . Variables reflect the entire
dataset (y i ). The ultimate goal of SVM is to detect an optimal server
hyperplane (wTx = y) that maximises the margin among two bounding
planes (wTx = γ + 1 and wTx = γ − 1), resulting in the smallest misclas-
sification error possible. The orientation vector w Rq is stated to be
normal to the bounding planes in SVM, as is the bias, which specifies
the location on the sever plane from the origin. The planes wTx = γ + 1
and wTx = γ − 1 are discovered to be included within the sets A + and
A-, respectively, the margin for the border planes.
SVM is developed based on the following model:
1
minimise(w, s) wr w + Ce T s (1)
2
subject to D(Aw − eγ ) + s ≥ e with s ≥ 0 (2)
To optimally separate the hyperplanes using support vectors, Eqs. (1)

and (2) are considered a two optimization issue with a solution at (w, γ ).
The server hyperplane, which has length q and is given by, For any sample
vector under consideration, it acts as a classifier.
⎧
⎨ > 0, then xc A+
xt w − γ < 0, then xc A− (3)
⎩
= 0, then xc A + or xc A−
3.2 Hybrid Support Vector Machine (SVM) with Grey Wolf

Optimization (GWO)
The data points are categorised in SVM stationed to their closeness
to one of two parallel hyperplanes that distinguish datasets relationship
to two various class memberships. SVM’s classification performance is
comparable to that of a straight forward solution technique in terms of
accuracy. The suggested hybrid SVM–GWO classifiers for data classifica-
tion in this research contribution combine the grey wolf optimization
(GWO) proposed in this section with the linear and non-linear SVM clas-
sifiers discussed in Sect. 3.2. The proposed system design for SVM–GWO
classification in each phase is depicted in Fig. 4.
For selecting the feature sets of the database under consideration,
this proposed methodology integrates the concepts of SVM and GWO.
The proposed self-regulated learning GWO-based SVM classifier allows
users to select features from datasets to improve generalisation and condi-
tioning of linear and non-linear classifiers. The following are the steps in
the suggested method:
Step 1: Set up the SVM module’s proximal planes and bias. Select
the kernel functions to be used for categorization as well. Fix the search
space’s D dimensions.
Step 2: The output proximal hyperplanes for each member of the
group are located at the SVM classifier. The training dataset is used to
derive the SVM parameters ‘C’ and ‘’.
Step 3: Determine the orientation vector and separation plane location
from the origin for each linear and non-linear SVM classifier.
Step 4: Now, as suggested, invoke self-regulated learning GWO.
Step 5: The fitness of individual member, in measured by the mean
square error (MSE), is then assessed as follows:
1 Σ 2 1 Σ( i )2
N
MSE = Ei − yk − dki . (4)
N N
I =1 i=1
The difference between the actual and desired outputs of the kth
output neuron in the ith sample is denoted by yk and dk, where N
is the number of training samples. The fitness function f is defined by
the MSE in this way (x). To avoid overfitting the classifier model, each
member’s fitness is assessed using the mean square error (MSE) on only
the validation set, rather than the whole training set.
Start
Five datasets (Medical and Non-medical) from UCI Machine

Learning Repository
Initialize proposed Test datasets

Training dataset
GWO parameters
Hybrid GWO
Initialize Based SVM
Population Training
Evaluate fitness function
Update particle velocity

Check and Check and and position,
update pbest update gbest
Stopping
criteria
Reached?
Trained SVM with optimal features and parameters selected
Perform Data Classification and diagnose the output class
Stop
Fig. 4 Flowchart of hybrid SVM–GWO classifier

Step 6: Based on the fitness, determine the SRL acceptability.

Step 7: Update each particle’s velocity and position equations.
Step 8: Stopping criteria—The algorithm continues stages 2–7 until
the maximum number of iterations is reached, or until a hard threshold
value is met. When the process is terminated, the values at which ideal
weights with the lowest MSE are discovered are reported.
Thus, the proposed hybrid GWO with SVM classifier computes the
best optimal features with the orientation vectors (‘w’ for linear classi-
fier and ‘v’ for non-linear classifier) and the location of the separating
plane from the origin “ so that the fitness reaches the minimum to
achieve better generalisation performance, while taking advantage of both
SVM and GWO classifier advantages. The proposed hybrid SVM-based
GWO classifier incorporates SVMfeatures into both linear and non-linear
GWO to compute the best features with the least MSE for successful data
classification.
4 Result and Discussion

PYTHON3.6IDE is used to implement the proposed methodology on
an Intel(R) Core (TP) i3-2410M CPU running at 3.20GHz with 8GB
RAM. This component is used to compare the selected reduct to prior
data knowledge. The suggested linear and non-linear hybrid SVM–GWO
classifiers’ learning performance is highlighted by the performance metrics
of classification accuracy, sensitivity, and specificity. From Tables 1, 2, 3
and 4, it can be deduced that the SVM–GWO outperformed other early
literature classifiers like the support ada boost method, vector machine,
random forest, also decision tree. Table 1 shows that the suggested
SVM–GWO classifier exceeded the results of conventional techniques,
indicating that it is more successful than previous classifiers. Each of the
five categories of datasets has been divided towards training also testing at
a 70-30 ratio. The ten-fold crosswise technique was used to validate the
training dataset. The training dataset is divided into ten equal subsamples
using the ten-fold crosswise procedure.
The analysis accuracy of a multivariate dataset is observed to be signifi-
cantly higher than that of previous techniques. In addition, the sensitivity
and specificity data that were successfully and wrongly classified are indi-
cated to 100, indicating a flawless classification rate. Looking at the
calculated simulation results for all of the other datasets in Table 1, the
performance of the SVM—GWO non-linear classifier is shown to be supe-

rior than the linear SVM classifier and other existing classifiers. Because
the gaussian distribution kernel in the SVM classifier explores itself in
effectively detecting the separation plane and conducting classification
based on the orientation vector, classification is better in the non-linear
scenario.
In support vector machine, random forest, ada boost, decision tree,
also proposed support vector machine (SVM) with grey wolf optimiza-
tion, the accuracy comparison for shuttle-2 vs 5, yeast-0-3-5-9 vs 7-8,
vowel0, cleveland-0 vs 4 and glass-0-1-4-6 vs 2 dataset in support vector
machine, random forest, ada boost, decision tree, and proposed support
vector machine (SVM) with grey (HSVMGWO). Figure 5 indicates that
the proposed system’s sensitivity outperforms all three current systems for
all five categories of datasets. Shuttle-2 vs 5 dataset has a 94.7% accuracy
rate, yeast-0-3-5-9 vs 7-8 has a 95.6% accuracy rate, vowel0 has a 96.3 %
accuracy rate, cleveland-0 vs 4 has a 98.4 % accuracy rate, and glass-0-1-
4-6 vs 2 has a 91.5 % accuracy rate. The total average accuracy was 95.3%,
but the other conventional approach only got to 90.28%.
Other existing systems achieve 91.3% SVM, 88.6% random forest,
76.8% ada boost, and 95.6% decision tree on the shuttle-2 vs 5 dataset,
which is lower than the proposed approach. The suggested system outper-
forms the yeast-0-3-5-9 vs 7-8 dataset by 87.4% of SVM, 79.4% of
random forest, 81.6% of ada boost, and 78.4% of decision tree. In the
vowel0 dataset, the suggested system outperforms 89.5% of SVM, 91.3%
of random forest, 90.5% of ada boost, and 81.6% of decision tree. 83.6%
Table 1 SVM–GWO accuracy comparison for conventional approaches
Accuracy comparison
Dataset Proposed Support Random ada boost decision
system vector forest tree
machine
shuttle-2_vs_5 94.7 91.3 88.6 76.8 95.6

yeast-0-3-5-9_vs_7-8 95.6 87.4 79.4 81.6 78.4
vowel0 96.3 89.5 91.3 90.5 81.6
cleveland-0_vs_4 98.4 83.6 98.5 86.5 78.6
glass-0-1-4-6_vs_2 91.5 91.3 93.6 88.6 81.5
Average 95.3 88.62 90.28 84.8 83.14
of SVM, 98.5% of random forest, 86.5% of ada boost, and 78.6% of deci-
sion tree in the cleveland-0 vs 4 dataset are lower than the proposed
system. The suggested system outperforms 91.3% of SVM, 93.6% of
random forest, 88.6% of ada boost, and 81.5% of decision tree on the
glass-0-1-4-6 vs 2 dataset.
and proposed support vector machine (SVM) with grey wolf optimiza-
tion, the sensitivity comparison for shuttle-2 vs 5, yeast-0-3-5-9 vs 7-8,
vowel0, cleveland-0 vs 4 and glass-0-1-4-6 vs 2 dataset is explained in
Table 2. (HSVMGWO). Figure 6 indicates that the proposed system’s
sensitivity outperforms all three current systems for all five categories of
datasets. when it comes to the proposed system shuttle-2 vs 5 dataset has a
sensitivity rate of 97.6%, yeast-0-3-5-9 vs 7-8 has a sensitivity rate of 95.6
%, vowel0 has a sensitivity rate of 97.3 %, cleveland-0 vs 4 has a sensi-
tivity rate of 99.4 %, and glass-0-1-4-6 vs 2 has a sensitivity rate of 95.5
%. The total average sensitivity was 97.08 %, although other traditional
approaches achieved maximum sensitivity.
Other existing systems achieve 97.6% SVM, 96.6 % random forest,
88.48 % ada boost, and 83.30 % decision tree on the shuttle-2 vs 5
dataset, which is lower than the proposed method. SVM, 89.58 % random
forest, 84.15% ada boost, and 88.21% decision tree are all lower than
Accuracycomparison
120
Accuracy in %
100 Proposed system

80
Support vector
60 machine
40 Random forest
20 ada boost
0
dataset
Fig. 5 SVM–GWO accuracy comparison for conventional approaches

Table 2 SVM–GWO sensitivity comparison for conventional approaches
Sensitivity comparison
machine
shuttle-2_vs_5 97.6 96.6 88.48 79.15 83.30

yeast-0-3-5-9_vs_7-8 95.6 92.54 89.58 84.15 88.21
vowel0 97.3 87.65 84.65 86.25 89.54
cleveland-0_vs_4 99.4 94.56 93.65 88.56 93.38
glass-0-1-4-6_vs_2 95.5 93.22 96.02 84.65 91.10
Average 97.08 92.914 90.476 84.552 89.106
the proposed system in the yeast-0-3-5-9 vs 7-8 dataset. In the vowel0

dataset, the suggested system outperforms 87.65% of SVM, 84.65% of
random forest, 86.25 % of ada boost, and 89.54 % of decision tree. The
suggested system outperforms 94.56% of SVM, 93.65% of random forest,
88.56% of ada boost, and 93.38% of decision tree on the cleveland-0 vs
4 dataset. SVM 96.02% random forest, 84.65% ada boost, and 91.10%
decision tree are all lower than the suggested system in the glass-0-1-4-6
vs 2 dataset (Fig. 6).
Sensitivity comparison
120
sensitivity in %
100 Proposed system

80 Support vector machine
60 Random forest
40
ada boost
20
decision tree
0
Dataset
Fig. 6 SVM–GWO sensitivity comparison for conventional approaches


and proposed support vector machine (SVM) with grey wolf optimiza-
tion, the specificity comparison for shuttle-2 vs 5, yeast-0-3-5-9 vs 7-8,
vowel0, cleveland-0 vs 4 and glass-0-1-4-6 vs 2 dataset in support vector
machine, random forest, ada boost, decision tree, and proposed support
vector machine (SVM) with grey (HSVMGWO). Figure 5 indicates that
the proposed system’s sensitivity outperforms all three current systems for
all five categories of datasets.Shuttle-2 vs 5 dataset has a specificity rate
of 91.6 %, yeast-0-3-5-9 vs 7-8 has a specificity rate of 98.6 %, vowel0
has a specificity rate of 96.3 %, cleveland-0 vs 4 has a specificity rate
of 97.4 %, and glass-0-1-4-6 vs 2 has a specificity rate of 94.5 %. The
proposed system achieved a total average specificity of 95.68 %, whereas
other conventional approaches only achieved a maximum of 92.6%.
In comparison, other current systems achieve 87.7% SVM, 82.12%
random forest,
78.84 % ada boost, and 74.4 % decision tree on the shuttle-2 vs 5
dataset, which is lower than the proposed method. The suggested system
outperforms 97.2 % of SVM, 96.5 % of random forest, 93.65 % of ada
boost, and 92.25% of decision tree in the yeast-0-3-5-9 vs 7-8 dataset. In
the vowel0 dataset, the proposed system outperforms 93.62 % of SVM,
94.65 % of random forest, 91.91% of ada boost, and 93.35 % of decision
tree. In the cleveland-0 vs 4 dataset, the suggested system outperforms
91.15 % of SVM, 89.15 % of random forest, 84.54 % of ada boost, and
86.54 % of decision tree.
Table 3 SVM–GWO specificity comparison for conventional approaches
Specificity comparison
Machine
shuttle-2_vs_5 91.6 87.7 82.12 78.84 74.4

yeast-0-3-5-9_vs_7-8 98.6 97.2 96.5 93.65 92.25
vowel0 96.3 93.62 94.65 91.91 93.35
cleveland-0_vs_4 97.4 91.15 89.15 84.54 86.54
glass-0-1-4-6_vs_2 94.5 93.65 95.5 90.14 87.15
Average 95.68 92.664 91.584 87.816 86.738
Specificity comparison
120
Specificity in % 100 Proposed system
60 Random forest
40
ada boost
20
0 decision tree
Dataset
Fig. 7 SVM–GWO specificity comparison for conventional approaches
The classification time period comparison in support vector machine,

random forest, ada boost, decision tree, and proposed support vector
machine (SVM) with grey wolf optimization for shuttle-2 vs 5, yeast-0-
3-5-9 vs 7-8, vowel0, cleveland-0 vs 4 and glass-0-1- 4-6 vs 2 dataset
in support vector machine, random forest, ada boost, decision tree,
and proposed support vector machine (SVM) with grey (HSVMGWO).
Figure 5 indicates that the proposed system’s sensitivity outperforms all
three current systems for all five categories of datasets. Shuttle-2 vs 5
dataset is 31.3M.sec, yeast-0-3-5-9 vs 7-8 is 13.31M.sec, vowel0 is 8.45
M.sec, cleveland-0 vs 4 is 7.13M.sec, and glass-0-1-4-6 vs 2 is 5.74M.sec
in terms of classification time duration.average total The time period for
five datasets was 13.1 milliseconds, but the usual approach only achieved
a minimum of 18.14 milliseconds.
In comparison, the suggested system achieves 41.14 M.sec of SVM,
46.45 M.sec of random forest, 50.51 M.sec of ada boost, and 39.14
M.sec of decision tree on the shuttle-2 vs 5 dataset. The suggested
system outperforms the yeast-0-3-5-9 vs 7-8 dataset by 25.35 M.sec of
SVM, 16.25 M.sec of random forest, 27.85 M.sec of ada boost, and
28.15 M.sec of decision tree. In the vowel0 dataset, SVM takes 7.95
milliseconds, random forest takes 8.21 milliseconds, ada boost takes 9.9
milliseconds, and decision tree takes 10.52 milliseconds. The suggested
system is faster than 14.52 M.sec of SVM, 13.25 M.sec of random
Table 4 SVM–GWO time period comparison for conventional approaches
Time duration comparison

machine
shuttle-2_vs_5 31.3 41.14 46.45 50.51 39.14

yeast-0-3-5-9_vs_7-8 13.31 25.35 16.25 27.85 28.15
vowel0 8.45 7.95 8.21 9.9 10.52
cleveland-0_vs_4 7.13 14.52 13.25 7.11 11.18
glass-0-1-4-6_vs_2 5.74 7.7 6.54 5.86 9.18
Average 13.186 19.332 18.14 20.246 19.634
forest, 7.11 M.sec of ada boost, and 11.18 M.sec of decision tree on
the cleveland-0 vs 4 dataset. The suggested system outperforms the glass-
0-1-4-6 vs 2 dataset by 7.7 M.sec of SVM, 6.54 M.sec of random forest,
5.86 M.sec of ada boost, and 9.18 M.sec of decision tree.
Tables 1, 2, 3 and 4 show the results of the hybrid support vector
machine with grey wolf optimization (HSVMGWO) for classification of
the selected five multivariate datasets. Reducets derived from forward
feature selection for vowel0, cleveland-0 vs 4, glass- 0-1-4-6 vs 2dataset
and backward feature removal approach for shuttle-2 vs 5, yeast-0-3-5-9
Time comparison
60
50 Proposed system
Time in Msec

30 Random forest
20 ada boost
10 decision tree
0
dataset
Fig. 8 SVM–GWO time period comparison for conventional approaches

vs 7-8dataset produce better performance results. The performance result

indicates that the built classifier may be utilised to classify the chosen
dataset in order to aid improved decision-making. Following the verifica-
tion of the results, it was demonstrated that planned study can be utilised
for real-time data categorization in a variety of contexts without error.
The system is capable of providing real-time solutions to difficulties, and
perfect achievement is achieved without delay.
5 Conclusion
This paper describes a hybrid support vector machine with grey wolf opti-
mization (HSVMGWO) technique for classifying five different types of
datasets. The suggested classifier’s simulated results in this chapter show
superior results, however, it should be highlighted that when used to solve
a complicated data classification problem, the HSVMGWO is easily stuck
in local optimization during the search process for effective features for
classification. This can be seen in the datasets shuttle-2 vs 5, yeast-0-3-
5-9 vs 7-8, vowel0, cleveland-0 vs 4, and glass-0-1-4-6 vs 2, where the
suggested algorithm got stuck with local minima numerous times and
had to be elevated to give classification results. In the future, knowledge
engineers may be able to create efficient decision support systems in real-
world scenarios employing hybrid classification methodologies including
two or more classifiers. Hybrid optimization techniques and bio-inspired
artificial intelligence approaches will generate stronger classifier models
in the future, which can be employed in the design and development of
decision support systems to increase efficiency.
References
Akay, M. F. (2009). Support vector machines combined with feature selection for
breast cancer diagnosis. Expert Systems with Applications, 36(2), 3240–3247.
Beckett, C., Eriksson, L., Johansson, E., & Wikström, C. (2017). Multivariate
data analysis (MVDA). Pharmaceutical quality by design: A practical approach,
201–225.
Chang, Y., Kim, N., Lee, Y., Lim, J., Seo, J. B., & Lee, Y. K. (2012). Fast and
efficient lung disease classification using hierarchical one-against-all support
vector machine and cost-sensitive feature selection. Computers in Biology and
Medicine, 42(12), 1157–1164.
Chen, Z., Lin, T., Tang, N., & Xia, X. (2016). A parallel genetic algorithm based
feature selection and parameter optimization for support vector machine.
Scientific programming, 2016.
Emamgholizadeh, S., & Mohammadi, B. (2021). New hybrid nature-based algo-
rithm to integration support vector machine for prediction of soil cation
exchange capacity. Soft computing, 25, 13451–13464. https://doi.org/10.
1007/s00500-021-06095-4
Galatenko, V. V., Lebedev, A. E., Nechaev, I. N., Shkurnikov, M. Y., Tonevitskii,
E. A., & Podolskii, V. E. (2014). On the construction of medical test systems
using greedy algorithm and support vector machine. Bulletin of Experimental
Biology and Medicine, 156(5), 706–709.
Gochhait, S. et al. (2021). Data interpretation and visualization of COVID-
19 cases using R programming. Informatics in Medicine Unlocked, 26(6),
Elsevier, ISSN: 0146–4116.
Gudadhe, M., Wankhade, K., & Dongre, S. (2010, September). Decision support
system for heart disease based on support vector machine and artificial neural
network. In 2010 International Conference on Computer and Communication
Technology (ICCCT) (pp. 741–745). IEEE.
Gürbüz, E., & Kılıç, E. (2014). A new adaptive support vector machine for
diagnosis of diseases. Expert Systems, 31(5), 389–397.
Kalimuthu, S. (2021). Sentiment analysis on social media for emotional predic-
tion during COVID-19 pandemic using efficient machine learning approach.
Computational intelligence and healthcare informatics, 215.
Kalimuthu, S., Naït-Abdesselam, F., & Jaishankar, B. (2021). Multimedia data
protection using hybridized crystal payload algorithm with chicken swarm
optimization. In Multidisciplinary approach to modern digital steganography
(pp. 235–257). IGI Global.
Karamizadeh, S., Abdullah, S. M., Halimi, M., Shayan, J., & javadRajabi, M.
(2014, September). Advantage and drawback of support vector machine func-
tionality. In 2014 international Conference on Computer, Communications,
and Control Technology (I4CT) (pp. 63–65). IEEE.
Kim, H. C., Pang, S., Je, H. M., Kim, D., & Bang, S. Y. (2003). Constructing
support vector machine ensemble. Pattern Recognition, 36(12), 2757–2767.
Kumar, R. S., & Arasu, G. T. (2015). Modified particle swarm optimization
based adaptive fuzzy k-modes clustering for heterogeneous medical databases.
Latha, P. (2014). SVM based automatic medical decision support system for
medical image. Journal of Theoretical & Applied Information Technology,
66(3).
Lee, S. H., Bang, M., Jung, K. H., & Yi, K. (2015, June). An efficient selec-
tion of HOG feature for SVM classification of vehicle. In 2015 International
Symposium on Consumer Electronics (ISCE) (pp. 1–2). IEEE.
Leng, Y., Sun, C., Xu, X., Yuan, Q., Xing, S., Wan, H., & …& Li, D. (2016).
Employing unlabeled data to improve the classification performance of SVM,
and its application in audio event classification. Knowledge-based systems, 98,
117–129.
Liu, R., Peng, J., Leng, Y., Lee, S., Panahi, M., Chen, W., Zhao, X. (2021)
Hybrids of support vector regression with grey wolf optimizer and firefly algo-
rithm for spatial prediction of landslide susceptibility. Remote Sensing, 13(24),
4966. https://doi.org/ https://doi.org/10.3390/rs13244966
Nguyen, T., Khosravi, A., Creighton, D., & Nahavandi, S. (2015a). Classifica-
tion of healthcare data using genetic fuzzy logic system and wavelets. Expert
Systems with Applications, 42(4), 2184–2197.
Nguyen, T., Khosravi, A., Creighton, D., & Nahavandi, S. (2015b). Medical data
classification using interval type-2 fuzzy logic system and wavelets. Applied soft
computing, 30, 812–822.
Orriols-Puig, A., & Bernadó-Mansilla, E. (2009). Evolutionary rule-based
systems for imbalanced data sets. Soft Computing, 13(3), 213–225.
Petrich, J., Gobert, C., Phoha, S., Nassar, A. R., & Reutzel, E. W. (2017,
August). Machine learning for defect detection for PBFAM using high reso-
lution layerwise imaging coupled with post-build CT scans. In Proceedings of
the 27th international Solid Freeform Fabrication Symposium.
Pham, B. T., Tien Bui, D., & Prakash, I. (2018). Bagging based support vector
machines for spatial prediction of landslides. Environment and earth science,
77 , 146.
Pourpanah, F., Lim, C. P., & Hao, Q. (2019). A reinforced fuzzy ARTMAP
model for data classification. International Journal of Machine Learning and
Cybernetics, 10(7), 1643–1655.
Purwar, A., & Singh, S. K. (2015). Hybrid prediction model with missing value
imputation for medical data. Expert Systems with Applications, 42(13), 5621–
5631.
Qinan, J., Lei, M., Jianfeng, H., QingQing, Y., & Jun, Z. (2014). A primary
study for cancer prognosis based on classification and regression using support
vector machine. In Frontier and future development of information technology
in medicine and education (pp. 909–920). Springer, Dordrecht.
Qiu, J., Wu, Q., Ding, G., Xu, Y., & Feng, S. (2016). A survey of machine
learning for big data processing. EURASIP Journal on Advances in Signal
Processing, 2016(1), 1–16.
Rokach, L. (2010). Ensemble-Based Classifiers. Artificial Intelligence Review,
33(1), 1–39.
Ros, G., Sellart, L., Materzynska, J., Vazquez, D., & Lopez, A. M. (2016). The
synthia dataset: A large collection of synthetic images for semantic segmenta-
tion of urban scenes. In Proceedings of the IEEE conference on computer vision
and pattern recognition (pp. 3234–3243).
Samanta, B., & Nataraj, C. (2009). Use of particle swarm optimization for
machinery fault detection. Engineering Applications of Artificial Intelligence,
22(2), 308–316.
Sheng, H., & Xiao, J. (2015). Electric vehicle state of charge estimation:
Nonlinear correlation and fuzzy support vector machine. Journal of Power
Sources, 281, 131–137.
Sindhiya, S., & Gunasundari, S. (2014, February). A survey on genetic algorithm
based feature selection for disease diagnosis system. In Proceedings of IEEE
International Conference on Computer Communication and Systems ICCCS14
(pp. 164–169). IEEE.
Stathopoulos, S., & Kalamboukis, T. (2015). Applying latent semantic analysis
to large-scale medical image databases. Computerized medical imaging and
graphics, 39, 27–34.
Vapnik, V., & Chapelle, O. (2000). Bounds on error expectation for support
vector machines. Neural Computation, 12(9), 2013–2036.
Wu, C. H., Chou, H. J., & Su, W. H. (2008). Direct transformation of coordi-
nates for GPS positioning using the techniques of genetic programming and
symbolic regression. Engineering Applications of Artificial Intelligence, 21(8),
1347–1359.
Zhan, Y., & Shen, D. (2005). Design efficient support vector machine for fast
classification. Pattern Recognition, 38(1), 157–161.
CHAPTER 8
Bioinformatics and Its Application

in Computing Biological Data
Sonali Patil and Annika Durve Gupta
1 Introduction
Bioinformatics is a study of interdisciplinary that creates software and
methods tools for analysing biological data, specifically complex and
large data sets. It is a study that combines biology, information engi-
neering, computer science, mathematics, statistics, and even recon-
struction, pattern recognition, simulation, machine learning, iterative
approaches, and molecular algorithms or modelling to analyse and inter-
pret biological data. For statistical and mathematical in silico analyses of
biological questions, bioinformatics has been used. One use of bioinfor-
matics is the examination of molecular sequences and genomic data. The
goal of bioinformatics, which is a combination of life sciences branches,
S. Patil (B)
Department of Bioanalytical Sciences, B. K. Birla College, Kalyan, Maharashtra,
India
A. D. Gupta
Department of Biotechnology, B. K. Birla College, Kalyan, Maharashtra, India

https://doi.org/10.1007/978-981-19-6506-7_8
134 S. PATIL AND A. D. GUPTA
is to develop methodologies and tools for organising, storing, systema-

tising, visualising, annotating, querying, understanding, and interpreting
large amounts of biological data, as well as managing large amounts of
data.
2 Goals
Bioinformatics encompasses both biological research that incorporates
computer programming and a collection of regularly used analysis “pipeli-
nes”, specifically in genomics field. Bioinformatics is commonly used to
identify candidate genes and single nucleotide polymorphisms (SNPs).
Such identification is usually done in order to better understand unique
adaptations, genetic basis of disease, attractive characteristics (in agri-
cultural fields), or population variances. Bioinformatics, often known
as proteomics, is an informal term for the study of the organisational
principles inside nucleic acid and protein sequences.
Gene discovery, drug design, sequence alignment, genome assembly,
gene expression, protein structure alignment, drug discovery, protein
structure prediction, and prediction of protein–protein interactions,
evolution modelling, genome-wide association studies, and cell divi-
sion/mitosis are just a few of the major research efforts in the field
(Frantzi et al., 2019).
Featured sub-disciplines within computational biology and bioinfor-
matics include:
● Implementing and creating computer programmes which allows

effective access to, utilisation and management various types of data
● Developing new algorithms (mathematical formulas) and statistical
measures for assessing relationships between members of large data
sets. Methods to find a gene within a sequence and/or predict
protein structure, as well as cluster protein sequences into families
of related sequences, are all available.
● Current research in molecular research and genomics techniques,
as well as advances in information technology, have combined to
create a large volume of molecular biology data over the previous
few decades. The word “bioinformatics” refers to the application
of mathematical and computer approaches to better understand
biological processes.
8 BIOINFORMATICS AND ITS APPLICATION IN COMPUTING … 135
Mapping and aligning DNA/protein sequences, analysing DNA and

protein sequences to compare them, and producing and displaying 3D
models of protein structures are all bioinformatics techniques (Benson
et al., 1998).
3 Relationship to Other Disciplines

Biological computation and bioinformatics are related but distinct disci-
plines of science. Sometimes, there is confusion between computational
biology and biology. Biological computation uses biology and bioengi-
neering to create biological computers, whereas bioinformatics uses
computation to learn more about biology. Computational biology and
bioinformatics are two fields that study the sequences of RNA, DNA,
and proteins in biological data. A major part of the Human Genome
Project and substantial advances in DNA sequencing technology resulted
in the rapid growth of the field of bioinformatics in the mid-1990s
(Allaby & Woodwark, 2004, Porter & Hajibabaei, 2021). Analysing
biological data and producing meaningful information that is used by
computer algorithms based on artificial intelligence, graph theory, data
mining, soft computing, computer simulation, and image processing.
It is based on theoretical foundations such as discrete system theory,
mathematics, control theory, statistics, and information theory (Cantor,
1998).
4 Databases
Bioinformatics provides different databases and tools for analysing biolog-
ical data.
4.1 The Use of Software and Tools

Bioinformatics software tools range from command-line utilities, i.e.
simple to even complex graphical applications also standalone web
services, all of which are offered from bioinformatics companies or
government agencies.
4.1.1 Bioinformatics Software that is Free and Open Source

Since the 1980s, a number of free and open-source software applica-
tions have gained popularity. All research groups have been able to
contribute to both bioinformatics and the wide range of open-source soft-

ware available like Bioconductor, BioPerl, Biopython, and many others by
combining new algorithms for determining the potential for innovative in
silico experiments, emerging types of biological readouts, and freely avail-
able open code bases. Community-supported plug-ins and idea incubators
are often used in commercial applications. Besides helping with de facto
standards and shared object models, they may also be able to assist with
the integration of biodata, which could be a big help in the future.
In order to facilitate the use of algorithms, data, and computing
resources on machines in different parts of the world, SOAP and REST-
based interfaces have been developed for many bioinformatics applica-
tions. There is a significant benefit in that end users are relieved of the
burden of maintaining software and databases.
The EBI (Biological Sequence Analysis) defines three types of basic
bioinformatics services: SSS (Sequence Search Services), MSA (Multiple
Sequence Alignment), and BSA (Fundamental Sequence Alignment). It
is possible to build standalone, web-based, or integrated bioinformatics
workflow management systems based on the use of standalone tools with
shared data formats.
4.1.2 Bioinformatics Workflow Management Systems

An application for managing bioinformatics workflows, commonly known
as workflows, refers to a collection of computational and data manip-
ulation procedures known as a workflow. It is intended to provide an
easy-to-use environment that allows individual application scientists to
easily create their own workflows within these systems;
● Make the process of sharing and reusing workflows among scientists

easier.
● Allow scientists to monitor the provenance of their workflows by
providing interactive tools enabling scientists to run workflows and
view outcomes in real-time.
Galaxy, UGENE, Kepler, Taverna, HIVE, and Anduril are some of the
platforms that provide this service.
5 Applications of Bioinformatics
Bioinformatics is being used in fields like microbial genome applications,
medicines, agriculture, and veterinary sciences.
5.1 Microbial Genome Applications

All genetic material from bacteria and eukaryotes, including chromosomal
and extrachromosomal components, is contained in microbial genomes.
This is a crucial field when it comes to Bioinformatics applications. Bioin-
formatics tools can help with DNA sequencing for usage in domains
including health and energy, in addition to analysing genome assembly
(Mann et al., 2021).
5.1.1 Waste Cleanup

For application in a range of sectors, millions of dangerous compounds
have been developed. These compounds are frequently discharged into
the environment as a result of human activities, contaminating water and
soil. Furthermore, many chemicals persist in the environment, causing
major health concerns to living organisms; it is vital that these toxins
be removed from the environment. The breakdown of xenobiotic or
chemicals substances by plants and bacteria is known as biodegradation
(Sadraeian & Molaee, 2009). Toxic compounds are degraded by biode-
grading bacteria through co-metabolism or mineralisation. Microbes
totally breakdown harmful substances by using them as carbon and energy
sources during the mineralisation phase, whereas poisonous compounds
are biotransformed into less toxic ones during co-metabolism. Micro-
bial remediation is an unique method of removing toxic substances
from the environment (Arora et al., 2009). Many microorganisms have
been isolated that can use harmful chemicals as their primary source
of energy and carbon, with many of them using microbial enzymes
like monooxygenases, dioxygenases, reductases, deaminases, and dehalo-
genases to break complicated chemical compounds down to carbon
dioxide and water. To improve bioremediation efficiency, the genes
encoding these enzymes have been found in a variety of microorgan-
isms and cloned into bacteria. The decomposition of a certain harmful
chemical necessitates the use of a specific microbe, which is dependent
on the chemical’s structure and the availability of enzyme systems in
bacteria. As a result, understanding chemicals (classification, distribution,
identification, toxicity, environmental features, and related dangers) as

well as their microbial biodegradation (xenobiotics degrading bacteria
genes, proteins, and enzymes) can help in bioremediation. Bioinformatics
has been integrated into all fields of biological sciences, providing a
platform for researchers to construct valuable computational tools for
human and environmental well-being. Bioinformatics has been merged
with biodegradation throughout the last few decades, and several bioin-
formatics tools that are useful in the subject of biodegradation have
been created. Databases, prediction of biodegradation pathway, predic-
tion systems for chemical toxicity, and next-generation sequencing are
examples of these technologies (Greene, 2002). Biodegradative databases
collect information about chemical biodegradation, such as xenobiotic-
degrading microorganisms, toxic chemical metabolic breakdown path-
ways, enzymes, and genes involved in biodegradation. Biodegradation
Network-Molecular Biology Database (Bionemo), MetaCyc, The Univer-
sity of Minnesota Biocatalysis/Biodegradation Database (UM-BBD), an
OxDBase Database of biodegradative oxygenases, and BioCyc are among
these databases (Caspi et al., 2012).
5.1.2 Biotechnology
Global economic and social challenges are being addressed by advances
in molecular modelling, pharmaceutical discovery, disease characteri-
sation, forensics, clinical health care, and agriculture in the biotech-
nology industry. Bioinformatics has reached unprecedented heights
among the biological disciplines as a result of public trust in biotech-
nology and biotechnology’s advancement. Automatic gene identification,
genome sequencing, prediction of gene function, phylogeny, drug design
and development, protein structure prediction, vaccine development,
organism identification, comprehending genomic and gene complexity,
protein functionality, structure, and folding and other bioinformatics
applications exist to speed up research in the field of biotechnology. The
use of bioinformatics in research allows researchers to complete long-
term research projects quickly, such as genome mapping. The future
demands of biotechnology will also be met by bioinformatics inno-
vation. The role of bioinformatics in many biotechnology disciplines
has been discussed, including genomics, drug design, proteomics, and
environmental biotechnology.
Genomics
Genomics refers to the study of genes and their expression. This subject
generates a large amount of data regarding gene sequences, their inter-
relationships, and their functions. Bioinformatics plays a crucial role
in managing this massive amount of data. It is becoming easier and
easier to detect systemic functional behaviour, as more complete genome
sequences for more animals become available, through bioinformatics.
Thompson et al. (1994) assert that bioinformatics is critical in structural
genomics, nutritional genomics, and functional genomics.
Proteomics
The study of the function, structure, and interactions of proteins
produced by a tissue, cell, or organism is known as proteomics. It includes
methods in biochemistry, genetics, and molecular biology. Massive
volumes of data on protein profiles, protein activity patterns, protein–
protein interactions, and organelle compositions have been generated
using advanced biological techniques. This huge amount of data can be
managed and accessed using bioinformatics software, databases, and tools.
Image analysis of 2D gels, peptide mass fingerprinting, and fingerprinting
of peptide fragmentation are just a few of the techniques established in
the field of proteomics so far (Hanash, 2003).
Comparative Genomics
In comparative genomics, bioinformatics is used to determine the genetic
structural and functional relationships between various biological species.
Transcriptomics
Transcriptomics is the study of groupings of all messenger RNA molecules
in a cell (Marini et al., 2021). This is also known as Expression Profiling,
and it comprises utilising a DNA microarray to measure the level of
mRNA expression in a specific cell group. Microarray technology creates
thousands of data values in a single run, while a single experiment neces-
sitates hundreds of runs. To analyse such vast amounts of data, a variety
of software packages are used. For transcriptome analysis, bioinformatics
is used in this way, allow to determine mRNA expression levels. RNA-
sequencing (RNA-seq) has also been included in the transcriptomics
category (Eagles et al., 2021). The quantity and existence of RNA in
a sample at a certain moment are determined using next-generation
sequencing. It’s used to investigate how the cellular transcriptome evolves

throughout time.
Cheminformatics
Cheminformatics (chemical informatics) is the study of chemical
substance information storage, indexing, searching, retrieval, and appli-
cation. It comprises the logical organisation of chemical data in order
to make chemical structures, characteristics, and interactions more acces-
sible. It is theoretically possible to design a chemical with the required
properties, detect and structurally modify a natural product, and test its
therapeutic effectiveness using computer algorithms employing bioinfor-
matics. Cheminformatics analysis includes procedures like as grouping,
similarity searches, QSAR modelling, virtual screening, and others.
Gene Expression
Gene expression regulation allows researchers to use genetic data to
construct molecular technologies that is the basis of functional genomics,
and it that can count the number of genes that are currently being
transcribed in each cell at any given time (e.g. gene expression arrays).
5.1.3 Climate Change Studies

Because of the accelerated sea level rise, loss of sea ice, and more and
longer powerful heat waves, climate change is a global problem. Bioin-
formatics may be able to help solve this problem by sequencing microbial
genomes, which reduces carbon dioxide as well as other greenhouse gas
levels. This contributes significantly to the stabilisation of global climate
change. In the bioinformatics sector, more location-specific research is
needed, taking into account the microbes of particular region and their
ability to reduce CO2 (Sinha, 2015).
5.1.4 Bio-Weapon Creation

One of the most serious dangers to home and land security will
continue to be the deliberate deployment of traditional or combinatorial
bioweapons. In the domains of synthetic, molecular, and computational
biology, the misuse of dual-use and how-to procedures and techniques
could lower the technological hurdles to launching assaults, even for
small organisations or individuals. A variety of biodefense techniques are
being developed using bioinformatics. On the other hand, existing algo-

rithms have failed to transform pathogen genetic data into standardised
diagnoses, broad-spectrum medicines, or rational vaccine development.
Bioinformatics has a limited impact, despite its potential. More than
a dozen biodefense databases and information exchange platforms lack
interoperability and a common layer, restricting biodefense enterprises’
scalability and development. As a result, the development of compu-
tational biology applications must be prioritised in order to leverage
next-generation genome sequencing for forensic operations, medical
intelligence, mitigation, and biothreat awareness.
According to the European Bioinformatics Institute, scientists had
sequenced the genomes of 3139 viruses, 1016 plasmids, and 2167
bacteria as of December 2012, some of which are publicly available
on the internet. Scientists will soon be able to produce infections by
making synthetic viruses, synthetic genes, and possibly entirely new crea-
tures, thanks to the availability of full genomes and the afore-mentioned
breakthroughs in gene synthesis.
Furthermore, the exponential growth in computer capacity, along with
public access to genetic knowledge and biological equipment, as well as a
lack of government oversight, raises concerns about biowarfare from non-
military sources. According to the US government, terrorist networks
are recruiting scientists capable of producing bioweapons, which “has
prompted countries to be more transparent about their attempts to clamp
down on the threat of bioweapons”.
Alternative Energy Sources

One of the greatest difficulties of the twenty-first century is finding
a renewable and affordable source of energy. Plants have long been
employed as a renewable energy source, and they remain one of the
most promising prospects in this field. Plant biomass growth and quality
must be constantly enhanced in order for plants to continue to provide
a cost-effective and renewable source of energy. Recent advancements
in genomics, aided by the advent of high-throughput sequencing and
genotyping tools, have opened up new options for plant breeding and
biotechnology development. The continual development of computers
and bioinformatics tools is required to analyse the large amounts of data
generated by these current genomics platforms. To employ genomics
for finding alternative energy sources, bioinformatics methods for gene
expression analysis with RNA-seq and for SNP genotyping can be applied.
It will make it possible. Engineering strategies are anticipated by the avail-

ability of large sequenced genomes, metabolic pathway reconstruction,
functional genomics investigations, the development of in silico models
at the genome scale, and synthetic biology approaches.
Renewable energy sources have a lot of potential in terms of biofuels.
Bioinformatics is required to understand and analyse biofuel manufac-
turing processes. Genetically engineered microalgal strains with suitable
lipid content have been made possible with recent advances in algal
genomics and other “omics” approaches.
Bioinformatics can also be used to discover alternative energy sources
in bacterial organisms. Bioinformatics is being used to examine the
genome of Chlorobium tepidum, a bacteria with an extremely large
genome.
5.2 Medicines
In medicine, bioinformatics has several uses, including gene research,
medicinal development, and prevention. Medical applications of bioin-
formatics are:
● Pharmaceuticals: Pharmaceutics research has relied heavily on

bioinformatics researchers, particularly in the field of infectious
diseases. Personalised medical research has also advanced thanks
to bioinformatics, with new therapies based on a person’s genetic
profile being discovered.
● Prevention: By analysing disease patterns, community healthcare
infrastructure, illness causes, and so on, bioinformatics, like pharma-
ceuticals, can be integrated into preventative medicine to produce
preventative medicine.
● Therapy: Bioinformatics can be effective in gene therapy, particu-
larly for single genes that have been harmed. Genetics scientists have
studied this application of bioinformatics and discovered that using
bioinformatics, one’s genetic profile can be improved.
5.2.1 Molecular Medicine

In modern biology, bioinformatics is the study of two fundamental infor-
mation flows. The first is the transmission of genetic information from an
individual organism’s DNA to a population with qualities that are similar
to those of the same species. The second is the flow of experimental data
from observed biological phenomena to explanation models, which is
then followed by more tests to put the models to the test. The organ-
isation of DNA sequence and protein 3D structural data collections was
one of the first initiatives in bioinformatics in the 1960s and 1970s. With
the development of biological investigations that produce vast volumes of
data quickly, it has grown into a thriving academic and corporate sector
(such as the multiple genome sequencing projects, the large-scale anal-
ysis of gene expression, and the large-scale analysis of protein–protein
interactions). Clinical medicine (including clinical medical information
systems) has long been affected by basic biological science, and a new
generation of epidemiologic, prognostic, diagnostic, and therapeutic tools
is emerging. Over the next decade, bioinformatics activities that appear to
be solely focused on basic research are expected to become increasingly
relevant in clinical informatics. DNA sequence information and anno-
tations, for example, will become more widespread in medical records.
Clinical information systems will soon incorporate bioinformatics tech-
nologies established for research. The focus of genetic disorder research
is turning away from single genes and towards uncovering networks
of genes at cellular level, unravelling their intricate connections, and
establishing their role in disease. As a result, a new era of individually
personalised treatment will emerge. Bioinformatics will aid and guide
clinical researchers and molecular biologists in taking advantage of the
advantages of computational biology. Clinical research teams who can
seamlessly move from clinical practice to the laboratory bench to the use
of these powerful computational tools will be the most prolific in the
coming decades.
Personalised Medicine
The medicine is a sort of treatment that is personalised to each indi-
vidual’s genetic composition. Personalised medicine is a type of medical
care in which each patient’s treatment is individually adjusted to meet
their specific needs. It is conceivable since we are genetically diverse
from one another. There are two important keys in the concept. To
begin, medical research attempts highlight how personalised medicine
is. Shifting medicines focus from reaction to prevention, selecting the
optimal therapy, reducing the length and cost of clinical trials, lowering
the overall estimated cost of health care, and lowering the likelihood of
adverse drug reactions are all attempts (Zhang and Hong, 2015).
The second aspect is information technology’s rapid improvement,

which has led to the development of novel technologies for decoding
human genomes, large-scale genetic variation research, and medical infor-
matics. Due to rapid development of new sequencing technologies,
sequencing has moved to a new level. Various sequencing techniques
have arisen since 2003, when human genome project was completed.
As a result, whole genome sequencing is now less expensive, reducing
from $2.3 billion in 2003 to just $1000 in 2016, and this trend may
continue in the next years or decades. Artificial intelligence (AI) and the
development of more powerful algorithms to produce learning machines
play a role in efficiently managing data sets and accurately projecting
outcomes. AI has already aided in the development of medical tech-
nology, from medication development to the development of assisted
systems for clinics, as a tool that can replicate brain-based cognitive func-
tion. Breast cancer diagnosis is one important application of deep learning
based on this concept.
Personalised medicine uses bioinformatics to analyse data from genome
sequencing or microarray gene expression studies to detect mutations that
may affect treatment response or prognosis. The Randomized Algorithm
and CADD can be useful tools (Cello et al., 2002).
Knowledge-based information is being incorporated into new algo-
rithms. SIFT is an evolutionary knowledge-based approach for predicting
mSNPs. SIFT uses a multiple sequence alignment between homolog
proteins to score the normalised probabilities for all possible substitutions,
whereas PolyPhen uses different sequence-based features and a position-
specific independent counts (PSICs) matrix from multiple sequence
alignment to predict the impact of mSNPs (Chenna et al., 2003). The
PANTHER algorithm uses a library of protein family to predict hazardous
mutations. 3D structural features can predict disease-related mSNPs.
Knowledge-based information has increased the predictability of algo-
rithms by more than 80%. For example, SNPs & GO is a functional
information-based technique that uses log-odd scores derived from anno-
tation in Gene Ontology (GO) words as input. MutPred assesses the
likelihood of structure and function gain or loss as a result of mutations
and identifies their impact.
5.2.2 Preventative Medicine

All physicians practise preventive medicine in order to keep their patients
healthy. It’s also a one-of-a-kind medical field recognised by the American
Board of Medical Specialties (ABMS). Preventive medicine focuses on

individuals, communities, and defined groups. To understand the patterns
and causes of diseases and health, it uses a variety of research methods,
such as bioinformatics, biostatistics, and epidemiology. As well as treating
obesity and blindness, it can also help with weight loss.
To gain a better understanding of the population’s health and disease
patterns and causes, as well as to integrate this information into disease-
prevention methods, research has been done. Multi-site, longitudinal
cohort studies are part of the research, and the faculty monitors a number
of investigator-initiated studies. Preventive medicine, often known as
preventive care, is a set of practices aimed at preventing diseases rather
than curing and treating their symptoms. Curative and palliative medicine,
as well as applied public health measures, can all be utilised to achieve this
goal.
The screening of neonates immediately after delivery for health issues,
such as metabolic disorders or genetic diseases, that are detectable but
not visible clinically in the newborn period, is an example of preventive
medicine.
Bioinformatics enables the collection and processing of data, as well
as the standardisation and harmonisation of data for scientific discovery
and the fusion of different data sources. Interoperability (the establish-
ment of an informatics system that allows access to and use of data from
many systems) will make scientific discoveries and vocations easier, as
well as potential for public health interventions. The National Cancer
Institute (NCI) has interoperable Cancer Biomedical Informatics Grid
(caBIG) which is the example of technologies used by population scien-
tists. Progress necessitates more than just tools. There are still issues, such
as a lack of common data standards, private data access hurdles, and chal-
lenges pooling data from different studies. To overcome these obstacles,
population scientists and informaticists are proposing new and creative
solutions.
To develop screening tests to diagnose diseases at an early
stage, researchers employ bioinformatics methods to analyse genomes,
proteomics, and metabolomics data.
The most recent example of bioinformatics-based preventative trials is
COVID-19 (Gochhait et al., 2021). The discovery of a large number of
coronaviruses, as well as coronavirus genomes that have been sequenced,
has proven to be an unusual occurrence for doing bioinformatics and
genomics studies on the virus family. The coronavirus is with large
genome when compared to other RNA viruses (26.4 kb to 31.7 kb).

Coronavirus has a high G + C content, ranging between 32 and 43
per cent. Based on their phylogenetic links, coronaviruses are divided
into three lineages: Alpha coronavirus, Beta coronavirus, and Gamma
coronavirus. Coronaviruses are classified into three lineages based on
their phylogenetic relationships: Alpha coronavirus, Beta coronavirus, and
Gamma coronavirus. As a result, the Beta coronavirus has been separated
into four subdivisions: A, C, B, and D.
Coronaviruses are well-known infections that can affect both animals
and humans. Between conserved genes and in the nucleocapsid gene’s
posterior, a varied number of small ORFs can be discovered. According
to these uses, bioinformatics studies have been involved in the preven-
tion of COVID-19. A large number of bioinformatics NGS experiments
are being conducted for COVID-19 research. Using bioinformatics tech-
niques and software, we can predict protein structure and determine
which genes are relevant to coronavirus.
5.2.3 Gene Therapy

The field of gene therapy involves implanting genetic components into
diseased cells to treat, cure, and prevent disease. The use of Bioin-
formatics in Gene Therapy includes identifying cancer types, analysing
protein targets, and assessing microRNA. The procedure of replacing a
patient’s damaged genes having functional one in their cells is known as
gene therapy. Gene therapy isn’t widely used since creating a generic gene
therapy strategy is challenging due to the fact that everyone’s genetic
makeup is different. Based on their DNA sequence, bioinformatics will
help determine the right gene target location for everyone.
5.2.4 Drug Development

Drug discovery is the most important application of bioinformatics.
Computational biology, a subset of bioinformatics, aids researchers in
deciphering disease causes and validating innovative, cost-effective treat-
ments. Infectious diseases are the greatest cause of death in children and
adolescents around the world. According to the World Health Orga-
nization, over 13 million people are being killed because of infectious
diseases each year. Molecular modelling and simulation can help speed
up the process of identifying therapeutic targets and screening drug
candidates, and more effective and safer drugs can be generated. In
the case of the COVID-19 epidemic, bioinformatics can be used to
develop a low-cost, effective medication (Imming et al., 2006, Sharma

et al., 2021). The Swiss Institute of Bioinformatics maintains http://cli
ck2drug.org/, which contains a comprehensive collection of tools, web
services, and databases, directly connected to drug discovery. These are
roughly divided into Thirteen categories: (1) databases, (2) molecular
modelling and simulation, (3) chemical structure representations, (4) the
structure of a protein is inferred via homology modelling, which is guided
by a homologue of known structure, and (5) Docking, (6) Prediction of
binding sites, (7) The structure of a protein can be inferred via homology
modelling, which uses a homologue of known structure as a guide, (8)
Drug Candidate Screening, (9) Prediction of Drug Targets, (10) Free
Energy Binding Estimation, (11) Ligand Design, (12) QSAR, and (13)
ADME Toxicity (Anderson, 2003). Many powerful and free software
products are funded by well-known institutions. These include databases
like ChEMBL and Swiss Sidechain, software tools like UCSF Chimera,
which is not only a 3D visualisation tool but also a platform for structural
biology software developers, Swiss Bioisostere for ligand design, Swiss
Similarity for virtual screening, Swiss SideChain, Swiss Target Prediction
to facilitate experiments that expand the protein repertoire by introducing
new proteins, and Swiss Similarity for virtual screening, Swiss Bioisostere
for PyMOL and CHARMM (Schrödinger) are examples of commercial
tools that often include free versions for students and teachers (Stoesser
et al., 1998).
5.2.5 Antibiotic Resistance

Antibiotic drug resistance is a major worry all over the world. Antibi-
otics put selective pressure on the propagation of resistance genes by
allowing bacterial isolates to exchange genetic material. Antibacterial
chemicals have altered the microbial populations of water, soil, and our
own microbiota as a result of their addition and overuse. We are power-
less in the face of antibiotic resistance. The development of new types
of medications is now a requirement of the period. Bioinformatics, on
the other hand, is revolutionising and intriguing the world of science
and technology (Muegge, 2003). Recent advancements in affordable
and quick DNA sequencing technologies have revolutionised diagnostic
microbiology and microbial surveillance. To date, there are at least 47
bioinformatics resources are freely available for detecting AMR deter-
minants in amino acids or DNA sequence data. These include CARD,
SRST2, ARG-ANNOT, Genefinder, MEGARes, ARIBA, AMRFinder,
KmerResistance, and ResFinder, among others. The type of accepted

input data, the presence/absence of software for searching within an
AMR determinant from other resources, and use of the search approach,
which can be based on mapping or alignment, are all factors that
differentiate bioinformatics resources. As a result, each technique has
strengths and limits in terms of AMR determinant detection sensitivity
and specificity, as well as application. The listed tools can be found at
public genomic data centres, downloaded from GitHub, or executed
locally. Both the European Nucleotide Archive (ENA) and the National
Center for Biotechnology Information (NCBI) provide online submis-
sion options for antimicrobial susceptibility sequencing and phenotypic
data, allowing other researchers to dive deeper into the data and test
new methodologies. Advances in whole genome sequencing, as well as
the use of online technology for real-time identification of AMR determi-
nants, are crucial in establishing control and preventative tactics to combat
the growing threat of AMR. The availability of technologies and DNA
sequence data is increasing, allowing for global disease surveillance also
genomics-based AMR tracking. Pipelines and databases must, however,
be standardised.
5.2.6 Evolutionary Studies

“Nothing in biology makes sense except in the light of evolution”,
stated Theodosius Dobzhansky, a brilliant American scientist. The study
of evolution is essential to understanding biological challenges and
improving the quality of life for humans. Kumar et al. (2008) used
bioinformatics to compare genomic data from many species.
Forensic Analysis
Biomolecular data is becoming increasingly important in forensic
research, and several European countries are building forensic databases
to preserve DNA profiles of known offenders’ crime sites and conduct
DNA testing. Statistical and technological developments, such as TFT
biosensors, DNA microarray sequencing, and machine learning algo-
rithms, which give an effective manner of organising and inferring
evidence, have strengthened the field (Bianchi & Lio, 2007). Nowadays,
homology modelling is employed to create 3D models in order to analyse
or justify our desired outcomes. Bioinformatics has changed the face of
molecular research by allowing researchers to determine gene structure or
sequence, protein structure, and molecular markers, as well as tie them to
other structures previously known. Bioinformatics research has provided

key methods for modelling a biological living cell system and docking
proteins, allowing scientists to develop viable therapeutic strategies to
tackle the growing problem of antibiotic resistance among infectious
illness victims. Bioinformatics is concerned with the analysis and interpre-
tation of a wide range of data, including nucleotide and protein domains,
amino acid sequences, protein structures, and the behaviour of protein–
ligand interactions at the molecular level. Homology modelling, often
known as computational biology, is a method of evaluating biological
annotations or data.
5.3 Agriculture
5.3.1 Development of Drought Resistant Varieties
Drought stress induced by unexpected precipitation is a huge danger to
global food supply, and its influence is only likely to grow as climate
change progresses. Understanding the impact of drought on crop and
plant responses is crucial for generating superior varieties with consis-
tent high yields to meet the growing food demand caused by a growing
population relying on diminishing land and water resources. The recent
introduction of innovative “-omics” technologies, like as proteomics,
genomics, and metabolomics, allows us to investigate and discover genetic
elements that underpin system complexity. The main challenge in this
genomics era is storing and managing the large amounts of data contained
in transcriptomics data or even genome scaffolds accessible for most of
the plant species; it is no exaggeration to claim that bioinformatics has
been well incorporated into modern-omics research. Sequence analysis
and de novo genome assembly tools, similarity searching tools, genome
sequencing tools, transcriptome, proteome, genome annotation tools,
and metabolome analysis, as well as visualisation tools, help us analyse
biological data and provide new insights into an organisation of biolog-
ical systems (Dahiya & Lata, 2017). This -omics knowledge might then
be applied to improve crop quality and production, as well as disease
resistance and abiotic stress tolerance. Bioinformatics is changing the
way molecular biology research is designed in the post-genomics age,
contributing significantly to scientific knowledge while also providing
new roles and perspectives to stress tolerance improvement genetic
engineering programmes.
5.3.2 Crop Improvement

As the climate changes and the world population expands, the strain on
our ability to produce enough food will increase. Fresh crop breeding
and adaptation of crops to new environment required to assure continued
food supply. Recent development in genomics has the potential to speed
up genetically based crop plant breeding. Linking genetic data to climate-
related agronomic characteristics for breeding objectives, on the other
hand, remains a substantial challenge that will require the collaboration
of a wide range of talents and knowledge. Bioinformatics and genomics
combination has potential to help security in food wrt climate change
by speeding the development of crops that are climate-ready. It makes
extensive use of proteomic, metabolomic, genetic, and agronomic crop
development to develop more powerful, drought-resistant, and insect-
resistant crops. Cattle quality and disease resistance will improve as a
result. Stress tolerance genes and alleles can be identified, which can
lead to the production of stress-tolerant cultivars. Many approaches have
been developed to study physiology, expression profiling, and compara-
tive genomics. All metabolic pathways, including the glucose production
process, are included in the KEGG database. Genes implicated in the
ABA production pathway are heavily used in the development of drought-
resistant cultivars. Researchers can use KEGG databases to find out which
genes are involved in carbohydrate and ABA production. When a pathway
is discovered, the genes are studied for their participation in it and in
development. There has been progress in cereal varieties development that
produce larger yields. These varieties will aid agriculture in flourishing in
areas with inadequate soil, enabling for the expansion of the worldwide
production base. Crop cultivars that can thrive in low-water conditions
are being produced as well. Plant science and industry have crossed the
genomics threshold with the completion of Arabidopsis thaliana genome
sequence and the preliminary sequence for genome of the rice.
5.3.3 Insect Resistance

Insecticide resistance is a significant concern for insect pest control
programmes in domains such as crop protection, human and animal
health, and so on. The proteins encoded from a specific class of insect
genes provide resistance to several pesticides. With the latest genome
sequencing, high-throughput genomics, and proteomics efforts on a
variety of insects, bioinformatics data processing approaches have become
increasingly important to aid scientists in interpreting new informa-

tion about the insects (Breton et al., 2021). As a result, bioinformatics
academics and experts have begun developing specialised databases and
tools for different industries. To comprehend the systems level physi-
ology, biology, host–pathogen interaction, disease mechanisms, growth
and development insect resistance, of numerous key insects, molec-
ular biologists use high-throughput genomes, transcriptomics, regulatory
genomics, epigenomics, and proteomics approaches. The massive amount
of data generated by these technologies necessitates a highly logical
mining and analysis of the entire data set, which may be accomplished
using well-established bioinformatics approaches and tools in the field.
5.3.4 Improve Nutritional Quality

Nutrigenetics is the study of how food molecules, genes, and gene
function interact, whereas nutrigenomics is the study of how dietary
molecules, genes, and gene function interact. If nutrition researchers
wish to be regarded a significant partner in the genetics and genomics
arena and take full advantage of the many new opportunities, they must
make a serious effort to include bioinformatics expertise into their tool-
boxes. Bioinformatics describes genomes, epigenomics, transcriptomics,
proteomics, and metabolomics, which are all important aspects of nutrige-
nomics. Bioinformatics is expanding in every field of biology, and it
has had a positive impact on agricultural development. By gathering,
preserving, analysing, and integrating vast amounts of metabolomics,
genomes, and proteomics data, bioinformatics allows users to efficiently
analyse large amounts of data. In order to improve crop nutritional value
and yield, bioinformatics makes data and tools available to anybody,
including individuals, companies, and industries. In addition, in silico
simulations can be used to detect complex interactions between protein–
protein, structures of model protein, and decipher the genetic and
physical high-resolution network present in plants.
5.4 Veterinary Science

Veterinary science study has advanced to a greater level because to bioin-
formatics. Bioinformatics is used in this subject to conduct sequencing
research on animals such as cows, pigs, and sheep. As a result, overall
productivity has increased, and cattle health has improved. Bioinfor-

matics has also benefited scientists in the creation of new vaccine target
identification methods (Pomerantsev et al., 1997).
6 Conclusion
Bioinformatics has become an important part of a variety of biological
fields. Bioinformatics methods such as signal processing and image enable
the extraction of conclusions that are useful from larger amounts of raw
data in experimental and molecular biology. In the realm of genetics, it
aids in the annotation and sequencing of genomes as well as their reported
mutations. Through biological literature, text mining, and the creation of
gene ontologies and biological, it aids in the organisation and querying of
biological data. It can also be used to find the expression and control of
proteins and genes. Bioinformatics tools help in the analysis, comparison,
and interpretation of genomic and genetic data, as well as knowing the
evolutionary elements in the molecular biology. It also aids in the inves-
tigation and cataloguing of biological pathways and networks on a more
integrated level, that are crucial aspects of systems biology. In structural
biology, it aids in the modelling and simulation of RNA, proteins, DNA,
and biomolecular interactions.
References
Allaby, R. G., & Woodwark, M. (2004). Phylogenetics in the bioinformatics
culture of understanding. Comparative and Functional Genomics, 5, 128–146.
Anderson, A. C. (2003). The process of structure-based drug design. Chem-
istry & Biology, 10, 787–797.
Arora, P. K., Kumar, M., Chauhan, A., Raghava, G. P., & Jain, R. K.
(2009). OxDBase: A database of oxygenases involved in biodegradation. BMC
Research Notes, 2, 67.
Benson, D. A, Boguski, M. S., Lipman, D. J., Ostell, J., & Ouellette, B. F.
(1998). GenBank. Nucleic Acids Research, 26(l), 1–7.
Bianchi, L., & Lio, P. (2007). Forensic DNA and bioinformatics. Briefings in
Bioinformatics, 8(2), 117–128.
Breton, G., Johansson, A. C. V., Sjödin, P., Schlebusch, C. M., & Jakobsson,
M. (2021). Comparison of sequencing data processing pipelines and appli-
cation to underrepresented African human populations. BMC Bioinformatics,
22(2021), 488. https://doi.org/10.1186/s12859-021-04407-x
Cantor, C. R. (1998). How will the Human Genome Project improve our quality
of life? Nature Biotechnology, 16(3), 212–213.
Caspi, R., Altman, T., Dreher, K., Fulcher, C. A., Subhraveti, P., Keseler, I. M.,
Kothari, A., Kubo, A., Krummenacker, M., Latendresse, M., Mueller, L. A.,
Ong, Q., Paley, S., Subhraveti, P., Weaver, D. S., Weerasinghe, D., Zhang,
P., & Karp, P. D. (2012). The MetaCyc database of metabolic pathways and
enzymes and the BioCyc collection of pathway/genome databases. Nucleic
Acids Research, 40(D1), D742–D753.
Cello, J., Paul, A. V., & Wimmer, E. (2002). Chemical synthesis of poliovirus
cDNA: Generation of infectious virus in the absence of natural template.
Science, 297 , 1016–1018.
Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T. J., Higgins, D. G., &
Thompson, J. D. (2003). Multiple sequence alignment with the Clustal series
of programs. Nucleic Acids Research, 31, 3497–3500.
Dahiya, B. L., & Lata, M. (2017). Bioinformatics impacts on medicine, microbial
genome and agriculture. Journal of Pharmacognosy and Phytochemistry., 6(4),
1938–1942.
Eagles, N. J., Burke, E. E., Leonard, J., et al. (2021). SPEAQeasy: A scal-
able pipeline for expression analysis and quantification for R/bioconductor-
powered RNA-seq analyses. BMC Bioinformatics, 22, 224. https://doi.org/
10.1186/s12859-021-04142-3
Frantzi, M., Latosinska, A., & Mischak, H. (2019). Proteomics in drug devel-
opment: The dawn of a new era? Proteomics Clinical Applications, 5,
e1800087.
Gochhait, S. et al. (2021). Data Interpretation and Visualization of COVID-
19 Cases using R Programming. Informatics in Medicine Unlocked, 26(6).
Elsevier. ISSN: 0146-4116.
Greene, N. (2002). Computer systems for the prediction of toxicity: An update.
Advanced Drug Delivery Reviews, 54(3), 417–431.
Hanash, S. (2003). Disease proteomics. Nature, 422, 226–232.
Imming, P., Sinning, C., & Meyer, A. (2006). Drugs, their targets and the nature
and number of drug targets. Nature Reviews Drug Discovery, 5, 821–834.
Kumar, S., Nei, M., Dudley, J., & Tamura, K. (2008). MEGA: A biologist-centric
software for evolutionary analysis of DNA and protein sequences. Briefings in
Bioinformatics, 9(4), 299–306.
Mann, L., Seibt, K. M., Weber, B., et al. (2021). ECCsplorer: A pipeline
to detect extrachromosomal circular DNA (eccDNA) from next-generation
sequencing data. BMC Bioinformatics, 23, 40. https://doi.org/10.1186/s12
859-021-04545-2
Marini, F., Ludt, A., Linke, J., & Strauch, K. (2021). GeneTonic: An
R/Bioconductor package for streamlining the interpretation of RNA-seq
data. BMC Bioinformatics, 22(610). https://doi.org/10.1186/s12859-021-

04461-5
Muegge, I. (2003). Selection criteria for drug-like compounds. Medicinal
Research Reviews, 23, 302–321.
Pomerantsev, A. P., Staritsin, N. A., Mockov, Y. V., & Marinin, L. I. (1997).
Expression of cereolysine ab genes in Bacillus anthracis vaccine strain ensures
protection against experimental hemolytic anthrax infection. Vaccine, 15,
1846–1850.
Porter, T. M., & Hajibabaei, M. (2021). Profile hidden Markov model sequence
analysis can help remove putative pseudogenes from DNA barcoding and
metabarcoding datasets. BMC Bioinformatics, 22, 256. https://doi.org/10.
1186/s12859-021-04180-x
Sadraeian, M., & Molaee, Z. (2009). Bioinformatics Analyses of Deinococcus
radiodurans in order to waste clean-up. In environmental and computer
science, 254. Second International Conference.
Sharma, A., Ghosh, D., Divekar, N., Gore, M., Gochhait, S., & Shireshi, S.
(2021). Comparing the socio-economic implications of the 1918 Spanish flu
and the COVID-19 pandemic in India: A systematic review of literature.
International Social Science Journal, 71, 23–36. https://doi.org/10.1111/
issj.12266
Sinha, S. (2015). Role of bioinformatics in climate change studies. J Science, 1,
1–9.
Stoesser, G., Moseley, M. A., Sleep, J., McGowran, M., Garcia-Pastor, M., &
Sterk, P. (1998). The EMBL nucleotide sequence database. Nucleic Acids
Research, 26(1), 8–15.
Thompson, J. D., Higgins, D. G., & Gibson, T. J. (1994). CLUSTAL W:
Improving the sensitivity of progressive multiple sequence alignment through
sequence weighting, position-specific gap penalties and weight matrix choice.
Nucleic Acids Research, 22, 4673–4680.
Zhang, L., & Hong, H. (2015). Genomic discoveries and personalized medicine
in neurological diseases. Pharmaceutics, 7 , 542–553.
Index
A B
Basic Local Alignment Search Tool
Ada boost algorithm, 111
(BLAST), 3, 7, 9, 12, 85, 88
Adenosine diphosphate (ADP), 92 Bioinformatics, 1–3, 5–7, 10, 12,
14–16, 25, 29, 38, 39, 41, 46,
Adenosine triphosphate (ATP), 91, 92
53, 57, 59, 64, 66, 68, 76,
Agriculture, 16, 63, 70, 137, 138, 83–89, 92–95, 105, 108,
149, 150 133–143, 145–147, 149–152
Algorithm, 3, 6, 7, 9, 10, 12, 14, 16,
20, 21, 23–25, 28, 31, 46, 51,
C
52, 69, 71, 72, 84, 86, 87, 93,
Cancer Biomedical Information Grid
112, 113, 116, 118, 122, 129,
(caBIG), 94, 145
133–136, 140, 144, 148
Cancer Genome Anatomy Project
Applications, 1, 3, 9, 10, 12, 14, 16, (CGAP), 94
20, 25, 28, 31, 37, 38, 42, 46, Catalogue of Somatic Mutation in
54, 57, 64, 66, 67, 69, 73, 75, Cancer (COSMIC), 95
84–86, 88–90, 92, 95, 134–138, Computational biology, 1, 2, 58, 64,
141, 142, 144, 146, 148 66, 74–77, 134, 135, 140, 143,
146, 149
Artificial intelligence (AI), 10, 16,
Convolutional neural network (CNN),
19–21, 23–33, 73, 106, 116,
26, 28–31
129, 135, 144
Corona virus disease (COVID), 95
Artificial neural networks (ANNs), 10, Cross linking mass spectrometry
21, 73, 116 (CXMS), 89
© The Editor(s) (if applicable) and The Author(s), under exclusive 155
license to Springer Nature Singapore Pte Ltd. 2022
https://doi.org/10.1007/978-981-19-6506-7
156 INDEX
D I
Database, 7, 14, 15, 38, 39, 48, 49, Information retrieval (IR), 101–104
59, 66–69, 76, 77, 84, 88, In silico analysis, 133
92–94, 104, 107, 112, 120, 135, International Cancer Genome
138, 139, 147, 148, 151 Consortium (ICGC), 57, 94
Decision tree, 115, 117, 122–128 International Nucleotide Sequence
Deoxyribonucleic acid (DNA), 2–5, 7, Database Collaboration
12, 14, 38, 46, 47, 53, 75, 84, (INSDC), 68, 94
85, 87, 92, 135, 139, 143, 147, Isotope coded affinity tag (ICAT), 87
148, 152
Difference gel electrophoresis
K
(DIGE), 87
Kinase Knowledge Base (KKB), 92
Discrete Probability Detector (DPD),
Kinase Pathway Database (KPD), 92
7
Kinase Sequence database (KSD), 92
DNA Data Bank of Japan (DDBJ),
Knowledge discovery (mining) in
68, 94
databases (KDD), 118
E L
Early Detection Research Network Liquid chromatography tandem mass
(EDRN), 93 spectrometry (LCMS), 87
Electronic health records (HER), 40 Literature, 7, 40, 58, 101, 103–108,
European Molecular Biology 122, 152
Laboratory (EMBL), 3, 68, 94
Expansion Terms (BET), 103 M
Expressed Sequence Tags (EST), 67, Machine learning (ML), 14, 16, 20,
69, 94 21, 23, 25, 31–33, 115, 116,
133, 148
Mass spectrometry (MS), 49, 56,
G 87–89
Gene ontology (GO), 76, 88, 144, Medicine, 12, 13, 16, 20, 40, 43,
152 45–49, 51, 53, 55–58, 84, 85,
Genetic Algorithms (GA), 72, 73, 95, 137, 141–145
115, 117 Micro RNA (MiRNA), 74, 92, 93
Grey wolf optimization (GWO), 112, Molecular biology, 2, 5, 7, 44, 107,
120, 122–124, 126–128 134, 139, 149, 152
Multivariate dataset, 122, 128
H N
Human Cancer Genome Project National Cancer Institute (NCI), 94,
(HCGP), 94 95, 145
INDEX 157
National Centre of Biotechnology S

Information (NCBI), 7, 67, 68, Software, 1, 6, 7, 10, 11, 55, 75, 76,
94, 148 84, 87, 89, 90, 133, 135, 139,
National Human Genome Research 147
Institute (NHGRI), 51, 95 Stable isotope labelling by amino acids
National Institutes of Health (NIH), in cell culture (SILAC), 87
41, 48 Support Vector Machine (SVM), 74,
Neural networks (NNs), 20–28, 33 112, 115–120, 122–129
New Expansion Terms (NET), 103 Systems biomedicine, 50, 56
Next generation sequencing (NGS),
38, 48, 57, 75, 93, 146
T
O Technology, 15, 31, 38, 40–44, 55,
Objective Digital Stains (ODSs), 7 57, 58, 64, 66, 69, 70, 72, 74,
75, 77, 83, 89, 92, 93, 111, 134,
138–140, 143–145, 148, 149,
P 151
Pest, 64–71, 73–77, 150 Translational bioinformatics (TBI),
Pesticide crystal protein (PCP), 73 39, 40, 45, 46, 50, 51, 57, 59
Plant ontology (PO), 76 Translational Stroke Research (TSR),
Post-translational modification (PTM), 42
38, 56 Two-dimensional gel electrophoresis
(2DE), 87
R
Random forest, 115, 122–128
Rectified linear unit (reLU), 26 U
Ribonucleic acid (RNA), 3–7, 12, 46, Unified Medical Language System
87, 92, 135, 139, 146, 152 (UMLS), 107

Bioinformatics

Uploaded by

Copyright:

Available Formats

Bioinformatics

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bioinformatics

Uploaded by

Copyright:

Available Formats

Edited by

ISBN 978-981-19-6505-0 ISBN 978-981-19-6506-7 (eBook)

7 Hybrid Support Vector Machine with Grey Wolf

About the Editors

Dr. Saikat Gochhait is Post-Doctoral Fellow from University of

engaged in various aspects of insect-host plant-natural enemy interactions

Dr. Saurov Mahanta is associated with National Institute of Electronics

Dr. Sunny Sharma is working with University of Jammu in the Depart-

© The Author(s), under exclusive license to Springer Nature 1

statistics to molecular biology and create computational & statistical tech-

Christian D. Wunsch established their dynamic programming algorithm

Basics of Bioinformatics: We know that all surviving entities are

C, A and G) in random sequence of three can give 64 possible combina-

human cell encloses 23 couples of chromosomes (separate threads of

3 Perspectives of Computer Science

database similarity search, motif detection, Markov chains or informa-

1. Recovery and exploration of sequence: Linear sequence of DNA,

FASTA is a text-based arrangement for representing either base

Humans Insulin gene sequence in FASTA format

Similarly, protein sequence in FASTA format for both proteins can be

Insulin protein in Humans in FASTA format

2. Graph theory: Graph theory, also called as graph methodology,

3. Artificial intelligence (AI): Here the bioinformatics functions

example of antifungal protein chitosanase/glucanase, soil bacterium

Drug development has benefitted a lot from simulation studies. It is

Image processing: It is essential and assists the biologists and scientists

used in breast cancer treatment, which works by binding on oestrogen

Forensic Analysis and Bioinformatics: Forensic analysis is largely

this is feasible with a sustainable agricultural production model. Here,

All these examples suggest the wide and ever-growing applications of

5 Conclusions and Future Prospects

Fares, H., DiNicolantonio, J. J., O’Keefe, J. H., & Lavie, C. J. (2016).

metabarcoding datasets. BMC Bioinformatics, 22, 256. https://doi.org/10.

Artificial Intelligence in Biological Sciences:

Uma Dutta, Nikhil Danny Babu, and Girish S. Setlur

© The Author(s), under exclusive license to Springer Nature 19

mate or hunting prey, etc. Even micro-organisms have been observed

technical details and broadly discuss the state-of-the-art applications of

2 Basics of Machine Learning (ML)

This encapsulates the basic structure of a neural network. Now we come

3 How Do Neural Networks Learn?

program. It requires a large dataset to train the neural network to produce

4 Applications of AI in the Life Sciences

4.1 Using Deep Neural Networks to Understand the Biological

hence underperforms when images of the same object in different posi-

neuroscientists excited about the prospect of using AI to understand

4.2 AI in Medical Diagnosis

cases. Researchers have already obtained positive results in recognizing

4.3 Decoding Protein Structure Using AI

obtained using experimental techniques like X-ray crystallography and

4.4 AI in Ecology and Conservation Biology

purpose and it can detect infrequent patterns and complex structures

training examples should be included and the complexity of the model

5.2 Adversarial Examples

5.3 Data Bias

A Review of Recent Advances

Chittaranjan Baruah, Bhabesh Deka, and Saurov Mahanta

© The Author(s), under exclusive license to Springer Nature 37

rapidly and efficiently. When it comes to productivity and translating

identification of prostate cancer biomarkers Prostate cancer diagnosis and