Computer-based analysis, visualization, and interpretation of antimicrobial peptide activities.

This chapter describes a computer-based method for analyzing the quantitative structure-activity relationships (QSAR) of antimicrobial peptides. Quantitative or qualitative activity measurements and known peptide sequences are used as input for the analysis. The analysis steps consist of the preprocessing which specifically deals with dilution series from an assay with luminescent bacteria, transformation of quantitative activity values into activity classes, a feature extraction method using molecular descriptors for amino acids, feature selection methods, visualization strategies, the classifier model design for discrimination of active and inactive peptides, and the in silico design of promising new peptide candidates.