Eigen Value ANalySis (EVANS) ‒ A Tool to Address Pharmacodynamic, Pharmacokinetic and Toxicity Issues

Abstract

Drug discovery is a continuously evolving area, essential for mankind albeit a very expensive process with a high attrition rate. The main challenge faced by pharmaceutical researchers today is to identify the major hurdles and validate the “developability” of a compound in the initial stages of drug development in order to select superior drug candidates with the best chances of success. This motivated us to introduce a “universal” approach for analyzing the pharmacodynamics, pharmacokinetics and toxicity profile of compounds based on the philosophy of QSAR. The present work deals with the development, validation and application of a novel QSPR formalism entitled EigenValue ANalySis (EVANS). This methodology encodes 3D structural information in terms of the atom pair distances along with molecular physicochemical properties to generate a set of unique hybrid descriptors, termed as “Eigenvalues.” The present article deals with the intricacies of the methodology and explores its applicability on a series of datasets for building pharmacodynamic QSAR models.

Application of the “EigenValue Analysis (EVANS)” Methodology to Build Quantitative Structure Pharmacokinetic Relationship Models

“…The maximum and minimum values of these computed distances then populate the upper and lower cells respectively of the distance matrix, thus generating a "max-min" distance matrix. This, in some way, accounts for the ensemble of structures that are likely In a previous study 7 , we have tested the EVANS methodology on pharmacodynamic datasets with encouraging results. This study focuses on the application of EVANS to build predictive QSPKR models using clinically derived PK data curated by Obach et al 29 The PK parameters modelled were VDss, CL, and t1/2, and models were built using various chemometric methods like MLR, RF, and SVM with linear and radial basis function kernels.…”

Application of the “EigenValue Analysis (EVANS)” Methodology to Build Quantitative Structure Pharmacokinetic Relationship Models

“…We have attempted to circumvent these problems by developing a QSPR methodology titled "Eigenvalue Analysis (EVANS)" that incorporates 3D structural information without the need to perform molecular alignment. We do so by projecting interatomic distances In a previous study 7 , we have tested the EVANS methodology on pharmacodynamic datasets with encouraging results. This study focuses on the application of EVANS to build predictive QSPKR models using clinically derived PK data curated by Obach et al 29 The PK parameters modelled were VDss, CL, and t1/2, and models were built using various chemometric To gauge the effectiveness of the EVANS formalism, we compare the validation metrics with the QSPKR models reported in the literature.…”

“…The steps in the EVANS methodology are shown in Figure 1. The formalism has been published earlier 7 , however, some aspects have been refined for use in QSPR modelling. We succinctly explain the workings of the EVANS formalism below.…”

“…The uniqueness of the EVANS methodology is that it incorporates 3D information and accounts for the contribution of multiple chiral states towards a particular biological endpoint without the need to perform an alignment procedure. This methodology has previously been benchmarked on pharmacodynamic datasets with encouraging results 7 . We now deploy the EVANS methodology to build predictive Quantitative Structure Pharmacokinetic Relationship (QSPKR) models using human intravenous clinical PK data.…”