Development of QSAR Models to Identify Mycobacterium tuberculosis enoyl-ACP-reductase Enzyme Inhibitors
bioRxiv – Bioinformatics
Pub Date : 2023-03-19
Aureo Andre Karolczak, Luis Fernando Saraiva Macedo Timmers, Rafael Andrade Caceres
Tuberculosis is a matter of global concern due to its prevalence in developing countries and the ability of mycobacterium to present resistance to existing therapeutic regimens. In this context, the present project proposes the QSAR (Quantitative Structure-Activity Relationships) modeling, as a way of identifying and evaluating, in silico, the estimated inhibitory activity of candidate molecules for the molecular improvement stages and/or in vitro assays, reducing research time and financial costs. For this purpose, the SAR (Structure-Activity Relationships) study conducted by He, Alian and Montellano (2007) was used on a series of arylamides, tested as inhibitors of the enzyme enoyl-ACP-reductase (InhA) of Mycobacterium tuberculosis . The Hansh-Fujita (classical) and CoMFA (Comparative Molecular Field Analysis) QSAR models were developed. The classic QSAR model obtained the best statistical result, using Multiple Linear Regression (MLR), with internal validation with correlation factor R2 = 0.9012 and predictive quality, according to the Stone-Geisser indicator Q2 = 0.8612. In the external validation, a correlation factor R2 = 0.9298 and a Q2= 0.720 was obtained, indicating a highly predictive mathematical model. The CoMFA Model managed to obtain a Q2 = 0.6520 in the internal validation, which allowed the energy fields around the molecules used to be estimated, this is essential information to foster molecular improvement. A library of small molecules was built, with analogs to those used in the SAR study, which were subjected to classic QSAR function, resulting in a group of ten molecules with high estimated biological activity. The molecular docking results suggest that the ten analogs identified by the classical QSAR model presented favorable estimated free energy of binding. The conclusion points to the QSAR methodology as an efficient and effective tool in the search and identification of promising drug-like molecules.
Read more here: Source link