Protein Sequences Features Extraction for Predicting Beta- Turns and their Types: A Review

Abstract

Beta-turns are considered to be important secondary structure types that have essential role in molecular recognition, protein folding and stability. They represent 25% of protein structures, therefore they are the most common type of non-repetitive or tight turns structures. Their prediction is considered to be an important issue in bioinformatics and molecular biology, because it provides valuable information and inputs for the fold recognition and drug design. There are many machine learning and statistical based approaches that were designed to predict beta-turns. Among the successful approaches that are based on machine learning are the approaches that used Neural Networks (NNs) and Support Vector Machines (SVMs) methods. These approaches used different features and features organizations. Among the most usable features in beta-turns prediction are the Position Specific Scoring Matrices (PSSMs) and the predicted secondary structure. This work gives a review of the most successful methods that are used for beta-turns prediction and the features as well as the organizations of these features that they used.