Abstract
Defensins are considered to play an important role in the innate immune system of virtually all life forms, from insects and plants to amphibians and mammals. They are classified into alpha, beta and theta-defensins. Fast and accurate computational prediction of defensin and defensin types will help in annotating unidentified defensin novel peptides. Identified defensins, owing to their small length and potent antimicrobial activity can be used effectively for development of new clinically applicable antibiotics. Thus predicting the defensin candidates will aid in accurate identification of novel peptide drugs. Support vector machines prediction model accuracy was 99% for defensin and defensin types. The results indicate that it is most accurate and efficient prediction method for defensin peptides. User friendly defensin web server is provided at www.defensinpred.cdac.in for the benefit of scientific community.
Keywords: Alpha defensin, beta defensin, compositional features, defensins, prediction server, support vector machine (SVM), immune system, novel peptides, antibiotics, Support vector machine.