Ensemble QSAR: A QSAR method based on conformational ensembles and metric descriptors

Abstract

Quantitative structure-activity relationship (QSAR) is the most versatile tool in computer-assisted molecular design. One conceptual drawback seen in QSAR approaches is the "one chemical-one structure-one parameter value" dogma where the model development is based on physicochemical description for a single molecular conformation, while ignoring the rest of the conformational space. It is well known that molecules have several low-energy conformations populated at physiological temperature, and each conformer makes a significant impact on associated properties such as biological activity. At the level of molecular interaction, the dynamics around the molecular structure is of prime essence rather than the average structure. As a step toward understanding the role of these discrete microscopic states in biological activity, we have put together a theoretically rigorous and computationally tractable formalism coined as eQSAR. In this approach, the biological activity is modeled as a function of physicochemical description for a selected set of low-energy conformers, rather than that's for a single lowest energy conformation. Eigenvalues derived from the "Physicochemical property integrated distance matrices" (PD-matrices) that encompass both 3D structure and physicochemical properties, have been used as descriptors; is a novel addition. eQSAR is validated on three peptide datasets and explicitly elaborated for bradykinin-potentiating peptides. The conformational ensembles were generated by a simple molecular dynamics and consensus dynamics approaches. The eQSAR models are statistically significant and possess the ability to select the most biologically relevant conformation(s) with the relevant physicochemical attributes that have the greatest meaning for description of the biological activity.

AVCpred: an integrated web server for prediction and design of antiviral compounds

“… The general dataset is comprised of below viruses with unique number of AVCs in brackets: Dengue virus 1, dengue virus 2, enterovirus, human adenovirus 5, human cox B1, human cox B5, human echovirus 13, human echovirus 9, human enterovirus 71, human enterovirus C, human polio virus 1, human rhinovirus, human rhinovirus 14, human rhinovirus 1B, human rhinovirus 2, human T lymphotropic virus, influenza A, influenza A (H1N1), influenza B, monkeypox virus, respiratory syncytial virus, Rift Valley fever virus (Cercopithecidae), sandfly fever Sicilian virus, SARS coronavirus, simian virus 40, Sindbis virus, vaccinia virus, vaccinia virus WR, variola virus, vesicular stomatitis virus, West Nile virus, yellow fever virus …”

“…To save time and money for discovering a new drug, researchers have widely used various computational methods to screen virtual libraries of compounds before the synthesis and animal testing of chemicals. Among the different approaches, quantitative structure–activity relationship (QSAR) is mostly used . In this approach, relationships connecting molecular descriptors and activity are used to predict the property of other molecules .…”

Newer QSAR Techniques

“…The method was developed by Pissurlenkar et al [8]. The FS-QSAR method was proved to have an effective predictive power compared to the traditional 2D-QSAR method due to the introduction of the similarity concept into the regression equation.…”

“…Among the newer QSAR techniques, hologram-based QSAR (HQSAR) has emerged as one of the powerful new two-dimensional (2D) fragment-based QSAR (FB-QSAR) techniques where specialized molecular fragments are employed. With the advancement of computation tools, new methods like Laborato´rio de QuimiometriaTeo´rica e Aplicada QSAR (LQTA-QSAR) [7], ensemble QSAR (eQSAR) [8], and other novel approaches like FB-QSAR, fragment-similarity-based QSAR (FS-QSAR), self-organizing map QSAR (SOM-QSAR), QUASAR (5D-QSAR), 6D-QSAR, and 7D-QSAR have been introduced by QSAR researchers with the intention of making the study of QSAR more useful, productive, and interpretable for designing new drug candidates with enhancement of the general acceptability of QSARs to the scientific community [9,10]. A number of recent publications have shown that HQSAR can give results comparable to complicated and exhaustive 3D-QSAR techniques, but the former is much easier to use [3,4].…”

Is conformation a fundamental descriptor in QSAR? A case for halogenated anesthetics

“…While molecular conformation in vacuum is governed by intramolecular interactions, the bioconformation is ruled by enzyme induced-fit; consequently, optimized and bioactive geometries are probably different to each other and, to obtain insight on the action mechanism of a drug and substituent effects, MD´s should not be generated over geometries optimized in a receptor-free environment. Efforts have been made to attenuate the drawback of using a conformation that is possibly wrong, e.g., by using average conformations, ensemble and multidimensional methods [5–8], but the risk of chemical–biological misinterpretation remains. Receptor-dependent QSAR methods have also been developed [9], but these are mostly complementary and are aimed at corroborating and/or rationalizing the results provided by the regression models, since the docking methodology itself provides intermolecular energies and docking scores that correlate with bioactivity.…”