TargetATPsite: A template-free method for ATP-binding sites prediction with residue evolution image sparse representation and classifier ensemble

• Published in: Journal of computational chemistry Vol. 34; no. 11; pp. 974 - 985
• Main Authors: Yu, Dong-Jun, Hu, Jun, Huang, Yan, Shen, Hong-Bin, Qi, Yong, Tang, Zhen-Min, Yang, Jing-Yu
• Format: Journal Article
• Language: English
• Published: Hoboken Wiley Subscription Services, Inc., A Wiley Company 30.04.2013; Wiley Subscription Services, Inc
• Subjects: Adenosine triphosphatase; Adenosine Triphosphate - chemistry; Binding Sites; classifier ensemble; Databases, Protein; Drug Design; Humans; Ligands; Molecular Dynamics Simulation; Protein Binding; protein functional annotation; protein-ATP binding sites prediction; Proteins; Proteins - chemistry; R&D; Research & development; residue evolution image; Software; sparse representation; Support Vector Machine; Thermodynamics
• ISSN: 0192-8651; 1096-987X; 1096-987X
• DOI: 10.1002/jcc.23219

Understanding the interactions between proteins and ligands is critical for protein function annotations and drug discovery. We report a new sequence‐based template‐free predictor (TargetATPsite) to identify the Adenosine‐5′‐triphosphate (ATP) binding sites with machine‐learning approaches. Two steps are implemented in TargetATPsite: binding residues and pockets predictions, respectively. To predict the binding residues, a novel image sparse representation technique is proposed to encode residue evolution information treated as the input features. An ensemble classifier constructed based on support vector machines (SVM) from multiple random under‐samplings is used as the prediction model, which is effective for dealing with imbalance phenomenon between the positive and negative training samples. Compared with the existing ATP‐specific sequence‐based predictors, TargetATPsite is featured by the second step of possessing the capability of further identifying the binding pockets from the predicted binding residues through a spatial clustering algorithm. Experimental results on three benchmark datasets demonstrate the efficacy of TargetATPsite. © 2013 Wiley Periodicals, Inc. Accurately localizing the protein‐ATP binding sites is important for protein function analysis and drug design. A template‐free machine learning‐based software named TargetATPsite is developed to predict protein‐ ATP binding residues and pockets from amino acid sequences. The binding residues are predicted by an ensemble classifier formed by support vector machines, and the input features are sparse representations of residue evolution images. Binding pockets are identified through spatially clustering predicted binding residues.