A Prediction Method for Transmembrane Segments in Proteins Utilizing Multiple Discrimination Functions

Daisuke Kihara [1] (kihara@kuicr.kyoto-u.ac.jp)
Toshio Shimizu [2] (slsimi@si.hirosaki-u.ac.jp)
Minoru Kanehisa [1] (kanehisa@kuicr.kyoto-u.ac.jp)

[1] Institute for Chemical Research, Kyoto University
Gokasho, Uji, Kyoto 611, Japan
[2] Department of Information Science,
Faculty of Science, Hirosaki University
3 Bunkyo-cho, Hirosaki 036, Japan


A novel method for detecting transmembrane segments in membrane proteins is presented. The method employs discriminant analysis to detect transmembrane regions in a query sequence by taking advantage of the information of different characteristics in different subgroups of membrane proteins classified according to the number of transmembrane segments. The method was applied to sequences of membrane proteins with experimentally determined locations of transmembrane segments.