In this article, we discuss recent advances and selected publications in the area of computer-based (in silico) metabolism and toxicity prediction, and comment on the opportunities for prediction of metabolism-based toxicity. The article also discusses the growing interest and importance of in silico prediction of metabolism and toxicity as tools to assist in library design and lead optimization. In addition, we consider the status of commercial metabolism and toxicity databases, and the need for future improvement, since data is often from older studies and from structures not typically considered as 'drug-like'. The article highlights the current status and potential of in silico models for property prediction, and the potential for linkage with vivo models to improve the integration of metabolism and toxicity into the drug discovery process and extrapolation to clinical studies. The article underscores that the future development, integration and application of in silico models will require a balance of local and global model approaches. The article also indicates that implementation and integration of models into drug discovery processes needs to be carried out in a rational and systematic manner, if we are to fully capitalize on the opportunity presented by in silico predictive modeling.

作者：Alan G E, Wilson；Anita C, White；Richard A, Mueller

来源：Current opinion in drug discovery & development 2003 年 6卷 1期