Predicting protein interaction sites from residue spatial sequence profile and evolution rate

Abstract

This paper proposes a novel method that can predict protein interaction sites in heterocomplexes using residue spatial sequence profile and evolution rate approaches. The former represents the information of multiple sequence alignments while the latter corresponds to a residueÕs evolutionary conservation score based on a phylogenetic tree. Three predictors using a support vector machines algorithm are constructed to predict whether a surface residue is a part of a protein-protein interface. The efficiency and the effectiveness of our proposed approach is verified by its better prediction performance compared with other models. The study is based on a non-redundant data set of heterodimers consisting of 69 protein chains.

Dataset reconstruction for protein interface identification using manifold learning method

“…Kini & Evans proposed predicting protein interaction sites by detecting the presence of "proline brackets" based on the observation that proline is the most common residue found in the flanking segments of interaction sites [14]. Also, amino acid composition has been used to prediction binding sites in many studies [3,[11][12][13][15][16][17][18]. Solvent accessibility was reported as the most discriminative feature in a predictor introduced by Porollo and Meller [19].…”

“…In recent years, many methods have been developed for studying protein interaction sites or location of interface residues [3][4][5][6][7][8][9][10][11][12][13][14]. Jones and Thornton analyzed the surfaces of protein complexes using a patch method and found the interfaces can be distinguished because it is typically large and hydrophobic and bury a large extent of non-polar surface area [7,8].…”

Computational Tools and Databases for the Study and Characterization of Protein Interactions

“…more rigid, to decrease the entropic cost upon complex formation (Cole and Warwicker 2002;Fleishman et al 2011). Sequence conservation has also proved to be a predictor (Lichtarge et al 1996;Wang et al 2006), although it remains a contentious issue as some works have shown that interfaces are not more conserved than the rest of the protein (Grishin and Phillips 1994;Caffrey et al 2004). Finally, it has been shown that interfaces are richer in -strands and long loops while -helical conformations are disfavoured (Neuvirth et al 2004).…”

A Sequence-Based Dynamic Ensemble Learning System for Protein Ligand-Binding Site Prediction

“…To evaluate the performance of our method, we adopted six evaluation measures: sensitivity (Sen), precision (Prec), F-measure (F1), specificity (Spe), accuracy (ACC), and Matthews correlation coefficient (MCC) [26], [35].…”