SVM-based method for the classification of protein sequence as secretory or non-secretory protein. It consists of different SVM modules based on different features of proteins such as compositions(Amino acid, physicochemical prperties, and dipeptide). In addition PSI-BLAST was also used to carry out similarity-based search. Finally a hybrid approach based SVM module was developed that encapsulates complete information of a protein sequence that is amino acid and dipeptide composiiton and PSI-BLAST. This module can classify the protein sequence between secretory and non-secrtory protein with an accuracy of 83%. Users have a choice to use any of these module for predcition of their query sequence.

Home  l  Submit  l  Algorithm  l  Help  l  Contact  

Institute of Microbial technology

If you are using this webserver, please cite:
Garg, A. and Raghava, G. P. S. (2008) A machine learning based method for the prediction of secretory proteins using amino acid composition, their order and similarity-search. In Silico Biology 8:129-140.
home submit query sequence About srtpred help suggestions or doubts