Premium
Multiple Machine Learning Based‐Chemoinformatics Models for Identification of Histone Acetyl Transferase Inhibitors
Author(s) -
Krishna Shagun,
Kumar Sushil,
Singh Deependra Kumar,
Lakra Amar Deep,
Banerjee Dibyendu,
Siddiqi Mohammad Imran
Publication year - 2018
Publication title -
molecular informatics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.481
H-Index - 68
eISSN - 1868-1751
pISSN - 1868-1743
DOI - 10.1002/minf.201700150
Subject(s) - cheminformatics , in silico , virtual screening , computational biology , histone , support vector machine , random forest , acetylation , docking (animal) , machine learning , naive bayes classifier , computer science , small molecule , chemistry , artificial intelligence , biology , biochemistry , pharmacophore , bioinformatics , gene , medicine , nursing
The histone acetyl transferase (HAT) are involved in acetylation of histones that lead to transcription activation in numerous gene regulatory mechanisms. There are very few GCN5 HAT inhibitors reported despite of their role in cancer progression. In this study, we have utilized in‐silico virtual screening approaches based on various machine learning algorithm to identify potent inhibitors of GCN5 HAT from commercially available Maybridge library. We have generated predictive chemoinformatics models based on k‐Nearest neighbour, naïve Bayesian, Random Forest and Support Vector Machine. Based on statistical parameters, the RF and SVM models have shown comparative performance. Therefore, we performed the virtual screening with these two models and the consensus hits were selected for further evaluation using molecular docking into the active site of GCN‐5 HAT. Finally, a set of 10 molecules were selected and subjected to biological evaluation. Subsequently, inhibition of acetylation shown by three out of the ten molecules in the in‐vitro experiments validated their utility as potential HAT inhibitors. Furthermore, the selected hits have also shown weak cell growth decrease in MCF‐7 cancer cell lines, which suggests that after subsequent structural optimization the identified molecules may further be explored for the development of anti‐cancer agents.