Improving decision making for drug candidates: A computational approach for benzthiazoles as antifungal

Abstract

To improve the decision making for selecting a molecule for synthesis from a set of virtually designed antifungal compounds, four <i style="">in silico</i> approaches are discussed and their optimization have been confirmed by <i style="">in vitro</i> methods. First, a set of compounds was designed over QSAR analysis and selected over predicted activities by best QSAR equation (r²>0.9) to increase the probability of finding new chemical entities. To reduce the burden of synthetic chemistry, Rule of Five was applied for screening druggable compounds. Second, OSIRIS was used to predict toxicity, mutagenicity and carcinogenicity. Third approach was prediction of biological activity spectra for substances (PASS) and fourth was molecular docking to analyze the actual interaction involved after binding with the target receptor. All compounds of the series passed rule of five. On the basis of OSIRIS prediction results, nine compounds were selected. PASS predictions have resulted into suitability of only one compound (test 30) as antifungal. Molecular docking at BioMed CAChe workstation was then performed for this compound. Comparison of active site residues at receptor binding and dock score with fluconazole has further confirmed the compound for synthesis. The synthesized compound has given MIC 16 µg/mL at antifungal assay against <i style="">Candida albicans</i> by potato dextrose agar method. It confirmed that <i style="">in silico</i> approaches are useful to find out a new drug with more accuracy.