A novel descriptor of amino acids and its application in peptide QSAR

A novel descriptor, vector of principal component scores (VSW) for weighted holistic invariant molecular index, was derived from the principal component analysis of a matrix of 99 weighted holistic invariant molecular indices of amino acids. The scale was then applied in three panels of peptide QSARs models constructed by partial least square (PLS). The correlative coefficient R(cum)(2) and the cross-validation correlative coefficient Q(LOO)(2) of three models were 0.861 and 0.835 for 58 angiotensin-converting enzyme inhibitors, 0.873 and 0.751 for 48 bitter tasting thresholds, 0.997 and 0.954 for 12 antimicrobial polypeptides, respectively. External validation was also performed to validate the predictive power of resulting models. Compared with other 2D or 3D descriptors, the VSW scales could better characterize structural features of peptides and provide more sound statistical models.