LAUSR

In this paper, four models of quantitative structure–activity relationship (QSAR) for 84 angiotensin I-converting enzyme inhibitory (ACEI) dipeptides, 169 ACEI tripeptides and 15 ACEI tetrapeptides and the total 268 peptides were established, respectively. During every modeling process, the structure of the peptides were firstly described using a set of amino acid descriptor SVHEHS proposed previously in our laboratory, then QSAR models were developed successfully by orthogonal signal correction combined with support vector machine (OSC-SVM). Furthermore, all models were validated by the reasonable statistical parameters, such as correlation coefficient (R2), cross-validated correlation coefficient (q2) and external validation coefficient (q2ext), etc. The results demonstrated that all developed QSAR models had excellent fitting accuracy and predictive ability. It could be a very promising approach for QSAR study on bioactive peptides using SVHEHS descriptor and OSC-SVM.