Mining Structural Patterns for Automatic Protein Function Prediction

Abstract

Protein function prediction is a challenging problem in bioinformatics. It has a great impact on the areas of diseases treatment and drug industry. Structure-based proteins representation plays an important role in proteins function prediction process. Three aspects of protein functions prediction have been considered: Predicting enzymes family and superfamily, classifying enzymes versus non-enzymes proteins and discriminating DNA-binding and non DNA-binding proteins. The thesis presents a modification to an existing protein representation approach which utilizes distance patterns between protein residues and a maximum cutoff. Also, the thesis presents a new protein structure representation for efficient protein function prediction. The new representation is based on three-dimensional patterns of protein residues. It utilizes atoms coordinates of protein residues, including the angles and distance patterns. This thesis also presents a study of the proposed protein representation and other protein-derived sequence, psychochemical and structure features to enhance the prediction of DNA-binding proteins and their classes.