A large number of compounds in the development stage got failure due to the unfavourable ADMET (absorption, distribution, metabolism, excretion and toxicity) profiles. The utilities of ADMET properties are becoming progressively more imperative in the drug discovery processes, assortment, development and promotion processes. In recent years several review papers have been published about the possibilities of the prediction or the ADMET properties using different structural features of the molecules, i.e., molecular descriptors, and utilizing multiple approaches. One of the most important approaches is QSAR modelling of the data derived from their activity profiles and their different structural features (i.e., quantitative molecular descriptors). More and more efforts are put to the field of ADMET predictions. This chapter will critically assess some of the most important and recently reported topics for the effective in silico predictions of the ADMET properties of the potential drug candidates based on QSAR modelling approaches.
Keywords: ADMET, molecular descriptor, QSAR modeling, logP, logD, artificial neural network, PSA, multiple linear regressions, machine learning, support vector machine, inductive logic programming
