Recent Advances in Chemometric Methodologies for QSAR Studies

Wei      Qi Lin; Jian      Hui Jiang; Hai      Long Wu; Guo-Li      Shen; Ru      Qin Yu

doi:10.2174/157340906778226418

Abstract

In recent years chemometrics has been undergoing an exciting development both in the applied areas and in the theoretical and methodological aspects. This review is focused on recent advances in chemometric methodologies for quantitative structure-activity relationship (QSAR) studies, and it covers multiple applications. QSAR is one of the tools for the computer-aided drug design; it is also an important branch of chemometrics. The feature or variable selection is an important aspect in QSAR studies. Basic requirements and different algorithms for feature or variable selection are briefly discussed. Moreover, an overview of the stateof- the-art chemometric methods developed to combat the shortcomings of conventional algorithms and their applications in QSAR is given. A survey of innovative chemometric approaches in QSAR model construction is also presented. Some remarks and outlook about QSAR studies applied to the computer-aided drug design have also been discussed.

Keywords: Computer-aided drug design, QSAR, chemometrics, variable selection, artificial neural networks, genetic algorithm, particle swarm optimization, support vector machine