A Neural Network Based Approach for GPCR Protein Prediction Using Pattern Discovery

Abstract

Artificial neural networks have been applied, as solutions, in several different problems within the field of bioinformatics. Similarly, pattern discovery algorithms have also been used to uncover hidden motifs in protein sequences, which have further contributed to understanding the problem of classification of different protein sequences. G-protein coupled receptors (GPCRs) represent one of the largest protein families in the Human Genome. Most of these receptors are major targets for drug discovery and development; therefore, they are of interest to the pharmaceutical industry. The technique used in this article combines both: neural network and pattern discovery methods to develop a protein prediction procedure in relation to its functional class, more specifically, to predict the GPCR protein. Vilo [2] proposed an algorithm to extract patterns of regular expressions from known GPCR protein sequences. Our contribution in this article is to combine these patterns as features for a neural network. We select patterns through the PCA (Principal Component Analysis) procedure and produce a learning machine for the prediction of the GPCR super class.