Abstract
A multi-scale parameterization approach, factor analysis scales of generalized amino acid information combined with auto cross covariance, was used to develop quantitative sequence-activity models of peptides using support vector machines. The results demonstrated that this approach could well characterize sequence features of the peptides studied.
Keywords: Factor analysis scales of generalized amino acid information (FASGAI), auto cross covariance (ACC), FASGAI-ACC, quantitative sequence-activity model (QSAM), support vector machines (SVM)