Importance of functional groups in predicting the activity of small molecule inhibitors for Bcl-2 and Bcl-xL.

Evasion of apoptosis owing to aberrant expression of Bcl-2 (B-cell lymphoma-2) anti-apoptotic proteins is a promising hallmark of cancer. These proteins are associated with resistance to chemotherapy and radiation. Currently available QSAR models are limited to a set of inhibitors corresponding to a particular chemical scaffold, and unified models are required to identify the differential specificity of diverse compounds toward inhibiting these targets. In this study, we predicted the factors driving differential activity and specificity implementing multiplexed QSAR analysis for a dataset of 1,649 reported inhibitors of Bcl-2 (B-cell lymphoma-2) and Bcl-xL (B-cell lymphoma-extra large). We developed QSAR models for seven diverse scaffolds and critically analyzed the chemical space with coupling factors. The correlation values of QSAR models for Bcl-2 and Bcl-xL range from 0.95 to 0.985. The MAE and sMAPE of the models were in the range of 0.052-5.4 nm and 0.41%-10%, respectively, signifying model robustness. The crucial descriptors and moieties accounting for the activity were benchmarked against experimentally determined binding patterns. The comprehensive analysis made in the study explores latent features of the chemical space in a broad perspective. Further, we have developed a user-friendly Web server for predicting a specific/dual inhibitor of Bcl-2 and Bcl-xL [http://www.iitm.ac.in/bioinfo/APPLE/].