Discovering New In Silico Tools for Antimicrobial Peptide Prediction

Marc      Torrent; M.      Victoria Nogues; Ester      Boix

doi:10.2174/138945012802002311

Abstract

Antimicrobial peptides (AMPs) are important effectors of the innate immune system and play a vital role in the prevention of infections. Due to the increased emergence of new antibiotic-resistant bacteria, new drugs are constantly under investigation. AMPs in particular are recognized as promising candidates because of their modularity and wide antimicrobial spectrum. However, the mechanisms of action of AMPs, as well as their structure-activity relationships, are not completely understood. AMPs display no conserved three-dimensional structure and poor sequence conservation, which hinders rational design. Several bioinformatics tools have been developed to generate new templates with appealing antimicrobial properties with the aim of finding highly active peptide compounds with low cytotoxicity. The current tools reviewed here allow for the prediction and design of new active peptides with reasonable accuracy. However, a reliable method to assess the antimicrobial activity of AMPs has not yet been developed. The standardization of procedures to experimentally evaluate the antimicrobial activity of AMPs, together with the constant growth of current well-established databases, may allow for the future development of new bioinformatics tools to accurately predict antimicrobial activity.

Keywords: Antimicrobial peptides, artificial neural networks, database, prediction, QSAR, support vector machines, peptide compounds, bioinformatic tools, antimicrobial spectrum