Abstract

Prediction of promiscuous coupling between G-protein coupling receptors (GPCRs) and various G-proteins is one of the challenges for current research. In this article, a novel method that based on the optimization of the coupling regions is proposed. The cross validation results demonstrate that it can achieve an average accuracy of over 80% for promiscuous coupling prediction. The results further suggest that the optimized coupling regions may be important for the processes of signal transduction pathways.

Keywords: G-protein coupled receptor, learning method, promiscuous coupling, signal transduction pathways, databases of Gproteins, intracellular regions, extracellular regions, Support Vector Machine, Performance Evaluation, Hidden Markov model