Computer Abetted Drug Design: Application of drug-target like protein prediction using learning algorithms of support vector machine

Abstract

In this paper, we explore the use of statistical learning approaches to predict drug-target like proteins from their primary sequences in order to facilitate the rapid discovery of new potential therapeutic targets from the large quantity of sequences in human genome. It was found that the Support Vector Machine (SVM) algorithm with a fine-tuned Gaussian kernel was able to make reasonably accurate prediction, which showed its potential to be used in the genome scale rapid drug target discovery, as a novel in silico approach supplementary to the conventional experimental approaches.