Computational Tools for ADMET Profiling
Denis Fourches, Antony John Williams, Grace Y. Patlewicz, Imran Hussain Shah, Chris Grulke, John F. Wambaugh, Ann Richard, Alexander Tropsha · Computational Toxicology · 2018
This chapter highlights some of the cheminformatics approaches and protocols that have demonstrated usefulness and high relevance for absorption, distribution, metabolism, excretion, and toxicology (ADMET) profiling in various published case studies and/or have attracted the attention of the research community. The ultimate success of the ADMET model depends on the accuracy of primary experimental data used for model development and the rigor of the computational tools and proper use of statistical model validation techniques. The chapter provides more details from the perspective of chemogenomics data curation. There have been several recent reviews outlining critical steps and best practices for developing rigorous and externally predictive QSAR models. A clear limitation of the hybrid chemical-biological modeling approach is that experimental high-throughput screening (HTS) data are required in order to use models for the prediction of in vivo effects for new chemicals.