Abstract

Outer membrane proteins (OMPs) play key roles in many cell functions and computational methods for discriminating OMPs from non-OMPs in genomic sequences are very meaningful. In this study, amino acid composition and auto covariance derived from PSI-BLAST profile, which could extract evolutionary information and sequence order effects effectively, were suggested to represent protein sequences for discriminating OMPs using support vector machine. To evaluate the performance of the proposed method, five-fold cross-validation tests were performed on three widely used benchmark datasets. Comparison with other previously developed methods showed that the proposed method was very competitive and may at least play an important complementary role to the existing methods.