About

SVMtop is a method for the prediction of transmembrane (TM) helices (the location of the TM helices) and topology (the orientation of the N-terminus) in alpha-helical membrane proteins. The method is based on support vector machines (SVM) in a hierarchical framework in which helix prediction is performed in the first stage, followed by topology prediction in the second. In the first stage, a SVM classifier is trained to distinguish the TM residues and the non-TM residues via incorporating several biological features of a TM helix in a lipid bilayer environment. The topology of the protein is deciphered in the second stage using the second SVM classifier trained to discriminate the non-TM residues into inside or outside loop residues. To achieve this goal, we developed a scoring function named Alternating Geometric Scoring Function (AGSF) based on the current understanding of topogenesis and membrane protein folding. The AGSF takes into account the inter-loop topogenic interactions and calculates the probability of the topology of the N-terminus loop located on the cytoplasmic or exoplasmic side. Standard benchmarks showed that SVMtop predicts about 70% of proteins with all helices and topology predicted correctly and less than 1% of false positive rate for identifying soluble proteins.

• Supporting information for Journal of Proteome Research

• Data sets for SVMtop

Reference

A. Lo, H.S. Chiu, T.Y. Sung, P.C Lyu, and W.L. Hsu.

Enhanced membrane protein topology prediction using a hierarchical classification method and a new scoring function, Journal of Proteome Research, 7, 2, 487 - 496, 2008.

If you use SVMtop, please cite.

>> Submit Sequence(s) Now !