Role of Amino Acid Properties to Discriminate the Transmembrane Protein Structure Segments

Membrane proteins play important roles in the biological process and accurate discrimination of membrane proteins from non-membrane proteins/globular ones would help to locate them in the genome sequences. However, the structural biology of membrane protein is limited due to the physiochemical complexities in determining its three dimensional structures. Thus the requirement of predicting membrane protein structure from sequence is increased and become a central problem in molecular biology; interestingly, several computational strategies and discriminating parameters were developed for the successful prediction of membrane protein structures. Studies have been reported that the transmembrane helical proteins could be discriminated with the accuracy of 90%, reflects the predation strength of the present algorithms. However, this accuracy fluctuates with other class of membrane proteins indicates the need for better physico-chemical observations of the specific folds. Thus, here performed a preliminary systematic analysis to study the role of various physico-chemicals, energetic and conformational amino acid properties to discriminate the transmembrane (TM) and nontransmembrane (NTM) segments of α and β class membrane proteins of diverse superfamily. The present study suggests the superfamily based discriminant properties to identify the transmembrane regions. We found that average numbers of surrounding residues, number of long-range contacts and total non-bonded energy can discriminate membrane proteins of all the superfamilies in the dataset with the accuracy of 85-91%. Thus we suggest the addition of these parameters can improve the accuracy of prediction for the specific superfamily.