In silico search of putative adverse drug reaction related proteins as a potential tool for facilitating drug adverse effect prediction.

Adverse drug reaction (ADR) is a significant issue in drug development and post-market applications. Different experimental and computational approaches need to be explored for predicting ADRs due to the complexity of their molecular mechanisms. One approach for predicting ADRs of a drug is to search for its interaction with ADR-related proteins (ADRRPs). In this work, this approach is tested on 11 marketed anti-HIV drugs covering protease inhibitors (PIs), nucleoside reverse transcriptase inhibitors (NRTIs), and non-nucleoside reverse transcriptase inhibitors (NNRTIs). An in silico drug target search method, INVDOCK, is used for searching the ADRRPs of each of these drugs. The corresponding ADRs of the predicted ADRRPs of each of these drugs are compared to clinically observed ADRs reported in the literature. It is found that 86-89% of the INVDOCK predicted ADRs of these drugs are consistent with the literature reported ADRs, and about 67-100% of the literature-reported ADRs of these drugs to various degrees is agreed with INVDOCK predictions. These results suggest that it is feasible to explore in silico ADRRP search methods for facilitating drug toxicity prediction.