Prediction of ADMET properties with machine learning: a trustworthy and explainable approach

Tariq, Zinnera

Prediction of ADMET properties with machine learning: a trustworthy and explainable approach

License

Citation

Tariq, Z. Prediction of ADMET properties with machine learning: a trustworthy and explainable approach. 2022. handle: http://hdl.handle.net/10230/55051

This citation was generated automatically.

Abstract

Efficacy and safety issues linked to absorption, distribution, metabolism, excretion (ADME) qualities, and different toxicities (T), are significant causes of drug development failures. Therefore, conducting an ADMET analysis on time is critical. The ability to predict these features quickly and accurately allows researchers to eliminate compounds that may present problems with ADMET. Given the current research and development model’s tremendous complexity, molecular modeling methodologies to find patterns in ADMET data and convert them into knowledge have been pursued. In this project, we have generated and validated models to predict ADMET properties and determine the realm of applicability domain and explainability of such models.