Computational modeling of P450s for toxicity prediction.

INTRODUCTION: Drug development is a time-consuming and cost-intensive process. On average, it takes around 12 - 15 years and approximately €800 billion to bring a new drug to the market. Despite introduction of combinatorial chemistry and establishment of high-throughput screening (HTS), the number of new drug entities is limited. In fact, a number of established drug entities have been withdrawn from the market because of drug-drug interactions (DDIs) and adverse drug reactions (ADRs). AREAS COVERED: This review covers the advancements in cytochrome P450 (CYP450) modeling using different computational/machine learning (ML) tools over the past decade. A computational model for identifying non-toxic drug molecule from the pool of small chemical molecules is always welcome in the drug industry. Any computational tool that identifies the toxic molecule at early stage reduces the economic burden by slashing the number of molecules to be screened. This review covers all issues related to CYP-mediated toxicity such as specificity, inhibition, induction and regioselectivity. EXPERT OPINION: Several computational methods for CYP-mediated toxicity are available, which are popular in computer-aided drug designing (CADD). These models may become helpful in toxicity prediction during early stages and can reduce high failure rates in preclinical and clinical trials. There is an urgent need to improve the accuracy, interpretability and confidence of the computation models used in drug discovery pathways.