[In-silico prediction of pharmacokinetic properties]

Abstract

In silico methods for predicting pharmacokinetic properties range from data-based approaches such as quantitative structure-activity relationships (QSARs), similarity searches, and 3-dimensional QSAR, to structure-based methods such as ligand-protein docking and pharmacophore modelling. Data-based modelling approaches are effective for many drug absorption, distribution, metabolism, and excretion (ADME) processes such as passive membrane permeation, where their molecular mechanism is barely delineated. Therefore QSAR approaches have been applied to simulate the relationships between ADME parameters and molecular structure and properties. In the present investigation, we describe the application of the genetic algorithm-combined partial least-squares (GA-PLS) method to QSAR modelling of various ADME properties. By selecting an appropriate set of molecular descriptors automatically using the genetic algorithm, many ADME properties could be well explained by simple molecular descriptors derived from the 2-dimensional chemical structure.