Cheminformatics driven exploration of chemical space for inhA inhibitors in Mycobacterium tuberculosis through virtual screening and molecular modeling

Abstract

The research work presented in this thesis comprises a computational investigation into the identification and optimization of novel inhibitors targeting the enoyl-acyl carrier protein reductase (InhA) enzyme of Mycobacterium tuberculosis, a key enzyme in mycolic acid biosynthesis and a validated anti-tubercular drug target. The study integrates cheminformatics, QSAR modeling, and structure-based drug design methodologies to explore the chemical space of known InhA inhibitors and to identify new bioactive scaffolds. A comprehensive activity landscape analysis was conducted to gain insights into structure— activity relationships (SAR) and detect activity cliffs among known inhibitors. A predictive QSAR model was developed using molecular descriptors and machine learning algorithms to classify compounds as active or inactive. This model was applied to screen selected phytochemicals derived from traditional Indian medicinal plants, and active compounds were subjected to molecular docking for further screening and to evaluate their binding potential. Structure-based virtual screening of the ZINC database was performed using molecular docking, followed by MD simulations, MM-PBSA free energy calculations, and DFT analysis to evaluate binding affinity, stability, and electronic properties. ADMET profiling was also carried out to assess drug-likeness. The findings highlight several potential lead compounds that exhibit strong inhibitory potential against InhA and warrant further experimental validation.