Authors: HongLiang Zou
Publish Date: 2014/08/09
Volume: 247, Issue: 11, Pages: 1141-1148
Abstract
Membrane protein is an important composition of cell membrane Given a membrane protein sequence how can we identify its types is very important because the type keeps a close correlation with its functions According to previous studies membrane protein can be divided into the following eight types singlepass type I singlepass type II singlepass type III singlepass type IV multipass lipidanchor GPIanchor peripheral membrane protein With the avalanche of newly found protein sequences in the postgenomic age it is urgent to develop an automatic and effective computational method to rapid and reliable prediction of the types of membrane proteins At present most of the existing methods were based on the assumption that one membrane protein only belongs to one type Actually a membrane protein may simultaneously exist at two or more different functional types In this study a new method by hybridizing the pseudo amino acid composition with multilabel algorithm called LIFT multilabel learning with labelspecific features was proposed to predict the functional types both singleplex and multiplex animal membrane proteins Experimental result on a stringent benchmark dataset of membrane proteins by jackknife test show that the absolutetrue obtained was 06342 indicating that our approach is quite promising It may become a useful highthrough tool or at least play a complementary role to the existing predictors in identifying functional types of membrane proteins
Keywords: