InterCriteria Analysis Implementation for Exploration of the Performance of Various Docking Scoring Functions

The present study describes an implementation of InterCriteria Analysis (ICrA) in the field of the computer-aided drug design and computational toxicology. ICrA strives to go beyond the nature of the criteria involved in a process of evaluation of multiple objects against multiple criteria, and, thus to discover some dependencies between the criteria themselves. The approach is based on the apparatus of the index matrices and the intuitionistic fuzzy sets. In this study new software capabilities, implemented in order to apply ICrA to in silico drug design, are presented. As a case study, ICrA is implemented to explore the performance of various scoring functions in docking. Docking, which is the most commonly used structure-based drug design method, has been applied to predict the binding mode and to provide a measure for the ligand binding affinity to the protein. In particular, ligands of the peroxisome proliferator-activated nuclear receptor gamma (PPAR\(\gamma \)), involved in the control of a number of physiologically significant processes, have been investigated towards prediction of their binding to the protein. A dataset of 160 tyrosine-based PPAR\(\gamma \) agonists with experimentally determined binding affinities has been used in this investigation. Docking combined with the in-house developed pharmacophore model as a filter has been applied. Various scoring functions and docking protocols have been tested in the molecular modelling platform MOE (v. 2019.01). ICrA has been applied to assess dependencies among the scoring functions. The analysis has demonstrated high positive consonance for two of the scoring functions – London dG and Alpha HB. None of the functions could be distinguished as a good predictor of the experimental binding affinity.