Identification Methods of G Protein-Coupled Receptors

Abstract

The G protein-coupled receptors (GPCRs) include one of the largest and most important families of multifunctional proteins known to molecular biology. They play a key role in cell signaling networks that regulate many physiological processes, such as vision, smell, taste, neurotransmission, secretion, immune responses, metabolism, and cell growth. These proteins are thus very important for understanding human physiology and they are involved in several diseases. Therefore, many efforts in pharmaceutical research are to understand their structures and functions, which is not an easy task, because although thousands GPCR sequences are known, many of them remain orphans. To remedy this, many methods have been developed using methods such as statistics, machine learning algorithms, and bio-inspired approaches. In this article, the authors review the approaches used to develop algorithms for classification GPCRs by trying to highlight the strengths and weaknesses of these different approaches and providing a comparison of their performances.

Bio-Inspired Data Mining for Optimizing GPCR Function Identification

“…For instance, GPCRPred (Bhasin & Raghava, 2004 ) based on SVM method for prediction of two levels for GPCRs families and sub-families. On the basis of the dipeptide composition from the primary amino acid sequence, PRED-GPCR (Papasaikas et al, 2004) provides a complement to the existing pattern database analysis servers and potentially a computational tool for GPCR family classification (Zekri et al 2011), GPCRTree is an online hierarchical classifications webserver (Davies et al, 2008). It is the first server to implement an alignment-independent representation of protein sequences and is also the first to classify sequences using a hierarchical classification, PCA-GPCR (Peng et al, 2010), and the most best web service is GPCR-MPredictor (Naveed and Khan 2012) that based on evolutionary approach and predicted GPCR at five level.…”