Quantitative Structure Activity Relationship

The goal of the present paper was to develop a quantitative structure-activity relationship (QSAR) method using a simple statistical approach, such as multiple linear regression (MLR) for predicting the blood-brain barrier (BBB) permeability of chemical compounds. The "best" MLR models, comprised log P and either molecular mass (M) or isolated atomic energy (E isol ), tested on a structurally diverse set of 66 compounds, is characterized the by correlation coefficients (R) around 0.8. The obtained models were validated using leave-one-out (LOO) cross-validation technique and the correlation coefficient of leave-one-out-R 2 LOO (Q 2 ) was at least 0.6. Analysis of a case from legal medicine demonstrated informative value of our QSAR model. To best authors' knowledge the present study is a first application of the developed QSAR models of BBB permeability to case from the legal medicine. Our data indicate that molecular energy-related descriptors, in combination with the well-known descriptors of lipophilicity may have a supportive value in predicting blood-brain distribution, which is of utmost importance in drug development and toxicological studies.

Quantitative structure ± activity relationships (QSARs) are frequently employed in medicinal chemistry projects, both to rationalize structure ± activity relationships (SAR) for known series of compounds and to help in the design of innovative structures endowed with desired pharmacological actions. As a difference from the so-called structure-based drug design tools, they do not require the knowledge of the biological target structure, but are based on the comparison of drug structural features, thus being defined ligand-based drug design tools. In the 3D-QSAR approach, structural descriptors are calculated from molecular models of the ligands, as interaction fields within a three-dimensional (3D) lattice of points surrounding the ligand structure. These descriptors are collected in a large X matrix, which is submitted to multivariate analysis to look for correlations with biological activity. Like for other QSARs, the reliability and usefulness of the correlation models depends on the validity of the assumptions and on the quality of the data. A careful selection of compounds and pharmacological data can improve the application of 3D-QSAR analysis in drug design. Some examples of the application of CoMFA and CoMSIA approaches to the SAR study and design of receptor or enzyme ligands is described, pointing the attention to the fields of melatonin receptor ligands and FAAH inhibitors.

The toxicity of polycyclic aromatic hydrocarbons (PAHs) was determined using mussel, sea-urchin and ascidian embryo-larval bioassays. Fluorescent light exposure enhanced phenanthrene, fluoranthene, pyrene and hydroxypyrene toxicity in comparison with dark conditions, but not naphthalene and fluorene toxicity. The toxicity of PAHs was inversely related to their K(OW) values following QSAR models derived for baseline toxicity of general narcotics, whereas the obtained regression using toxicity data from photoactivated PAHs significantly departed from the general narcosis model. Also, the mixture toxicity of five PAHs to the larval growth of the sea-urchin was compared with predictions derived from the concentration addition concept, indicating less than additive effects. Finally, we compared our toxicity data with worst-case environmental concentrations in order to provide a preliminary estimate of the risk to the marine environment. Naphthalene, fluorene and pyrene are not considered...

Quantitative structure-activity/property-activity relationships are developed that correlate the observed mutagenic behavior of 62 aminoazo derivatives and 12 of their reductive cleavage products with a variety of molecular descriptors calculated using quantum-chemical semiempirical methodology. Multilinear regression techniques using 8 descriptors are shown to account for more than 70% of the variation in the relative mutagenic activity of these compounds. Approaches using artificial neural networks in conjunction with fuzzy logic can account for about 95% of this variation using 8 descriptors. #

Quantitative structure−activity relationship modeling is one of the major computational tools employed in medicinal chemistry. However, throughout its entire history it has drawn both praise and criticism concerning its reliability, limitations, successes, and failures. In this paper, we discuss (i) the development and evolution of QSAR; (ii) the current trends, unsolved problems, and pressing challenges; and (iii) several novel and emerging applications of QSAR modeling. Throughout this discussion, we provide guidelines for QSAR development, validation, and application, which are summarized in best practices for building rigorously validated and externally predictive QSAR models. We hope that this Perspective will help communications between computational and experimental chemists toward collaborative development and use of QSAR models. We also believe that the guidelines presented here will help journal editors and reviewers apply more stringent scientific standards to manuscripts reporting new QSAR studies, as well as encourage the use of high quality, validated QSARs for regulatory decision making.

The estimation of accuracy and applicability of QSAR and QSPR models for biological and physicochemical properties represents a critical problem. The developed parameter of "distance to model" (DM) is defined as a metric of similarity between the training and test set compounds that have been subjected to QSAR/QSPR modeling. In our previous work, we demonstrated the utility and optimal performance of DM metrics that have been based on the standard deviation within an ensemble of QSAR models. The current study applies such analysis to 30 QSAR models for the Ames mutagenicity data set that were previously reported within the 2009 QSAR challenge. We demonstrate that the DMs based on an ensemble (consensus) model provide systematically better performance than other DMs. The presented approach identifies 30-60% of compounds having an accuracy of prediction similar to the interlaboratory accuracy of the Ames test, which is estimated to be 90%. Thus, the in silico predictions can be used to halve the cost of experimental measurements by providing a similar prediction accuracy. The developed model has been made publicly available at

In general perturbation methods starts with a known exact solution of a problem and add "small" variation terms in order to approach to a solution for a related problem without known exact solution. Perturbation theory has been widely used in almost all areas of science. Bhor's quantum model, Heisenberg's matrix mechanincs, Feyman diagrams, and Poincaré's chaos model or "butterfly effect" in complex systems are examples of perturbation theories. On the other hand, the study of Quantitative Structure-Property Relationships (QSPR) in molecular complex systems is an ideal area for the application of perturbation theory. There are several problems with exact experimental solutions (new chemical reactions, physicochemical properties, drug activity and distribution, metabolic networks, etc.) in public databases like CHEMBL. However, in all these cases, we have an even larger list of related problems without known solutions. We need to know the change in all these properties after a perturbation of initial boundary conditions. It means, when we test large sets of similar, but different, compounds and/or chemical reactions under the slightly different conditions (temperature, time, solvents, enzymes, assays, protein targets, tissues, partition systems, organisms, etc.). However, to the best of our knowledge, there is no QSPR general-purpose perturbation theory to solve this problem. In this work, firstly we review general aspects and applications of both perturbation theory and QSPR models. Secondly, we formulate a general-purpose perturbation theory for multiple-boundary QSPR problems. Last, we develop three new QSPR-Perturbation theory models. The first model classify correctly >100,000 pairs of intra-molecular carbolithiations with 75-95% of Accuracy (Ac), Sensitivity (Sn), and Specificity (Sp). The model predicts probabilities of variations in the yield and enantiomeric excess of reactions due to at least one perturbation in boundary conditions (solvent, temperature, temperature of addition, or time of reaction). The model also account for changes in chemical structure (connectivity structure and/or chirality paterns in substrate, product, electrophile agent, organolithium, and ligand of the asymmetric catalyst). The second model classifies more than 150,000 cases with 85-100% of Ac, Sn, and Sp. The data contains experimental shifts in up to 18 different pharmacological parameters determined in >3000 assays of ADMET (Absorption, Distribution, Metabolism, Elimination, and Toxicity) properties and/or interactions between 31723 drugs and 100 targets (metabolizing enzymes, drug transporters, or organisms). The third model classifies more than 260,000 cases of perturbations in the self-aggregation of drugs and surfactants to form micelles with Ac, Sn, and Sp of 94-95%. The model predicts changes in 8 physicochemical and/or thermodynamics output parameters (critic micelle concentration, aggregation number, degree of ionization, surface area, enthalpy, free energy, entropy, heat capacity) of self-aggregation due to perturbations. The perturbations refers to changes in initial temperature, solvent, salt, salt concentration, solvent, and/or structure of the anion or cation of more than 150 different drugs and surfactants. QSPR-Perturbation Theory models may be useful for multi-objective optimization of organic synthesis, physicochemical properties, biological activity, metabolism, and distribution profiles towards the design of new drugs, surfactants, asymmetric ligands for catalysts, and other materials.