Predicting the Activity of Antimicrobial Peptides with Amino Acid Topological Information

Abstract

In this paper, VSTPV, was recruited as a novel set of structural and topological descriptors derived from principal component analysis (PCA) on 85 structural and topological variables of 166 coded and non-coded amino acids. By using partial least squares (PLS), we applied VSTPV for the study of quantitative structure-activity models (QSARs) studies on two peptide panels as 101 synthetic cationic Antimicrobial polypeptides (CAMELs), and 28 bovine lactoferricin- (17–31)-pentadecapeptides (LFB). The results of QSARs models were superior to that of the earlier studies, with squared correlative coefficient R2 and cross-validated Q2 of 0.783, 0.656; and 0.864, 0.793, respectively. So, VSTPV descriptors were confirmed to be competent to extract information on 85 structural variables and to relate with biological activities.