| dc.description.abstract | Background: G protein-coupled receptors (GPCRs) are a large superfamily of membrane proteins and because of the difficulties in experimentally determining their structures, computational approaches are essential. Objective: GPCRTOP v.1.0 is an HMM-based web server which has been developed for predicting helical transmembrane (TM) segments and identifying GPCRs based on amino acid distribution patterns. The performance of the method was evaluated in comparison to other general TM prediction methodologies. Methods: 49093 unannotated human protein sequences were retrieved from TrEMBL-SwissProt. The InterPro database was used for finding the GPCR sequences in common with those predicted by GPCRTOP v.1.0. For those which were not in common, ten well-known TM predictors were utilized to analyse these sequences. Results: The results showed that 199 sequences were predicted as GPCRs by GPCRTOP v.1.0 whereas, there were 182 GPCR sequences in InterPro database. Among these sequences, 104 sequences were identified as GPCR by both GPCRTOP v.1.0 and InterPro database. The remaining sequences were then predicted by general TM predictors and their results showed 11.1% more agreement to that of GPCRTOP v.1.0 than InterPro database. Conclusion: GPCRTOP v.1.0 is useful for identifying GPCRs and determining their topologies with overall accuracy of similar to 99%. Here, we also announce the web availability of GPCRTOP v.1.0 (http://gpcrtop.tbzmed.ac.ir/services.aspx) and also describe its prediction features, which include protein type (i.e., GPCR or non-GPCR), number of TM segments, as well as the topology of the predicted GPCR. | |
