Predicting G-Protein-Coupled Receptor classes based on adaptive K-nearest neighbor algorithm

G-Protein-Coupled Receptor (GPCRs) play a key role in cellular signaling networks that regulate various physiological processes. The functions of many of GPCRs are unknown, because they are difficult to crystallize and most of them will not dissolve in normal solvents. This difficulty has motivated and challenged the development of a computational method which can predict the classification of the families and subfamilies of GPCRs based on their primary sequence so as to help us classify drugs. In this paper the adaptive K-nearest neighbor algorithm and protein cellular automata image (CAI) is introduced. Based on the CAI, the complexity measure factors derived from each of the protein sequences concerned are adopted for its Pseudo amino acid composition. GPCRs were categorized into nine subtypes. The overall success rate in identifying GPCRs among their nine family classes was about 83.5%. The high success rate suggests that the adaptive k-nearest neighbor algorithm and protein CAI holds very high potential to become a useful tool for understanding the actions of drugs that target GPCRs and designing new medications with fewer side effects and greater efficacy.