Prediction of long-range contacts in protein tertiary structures

• Main Author: Soh, Lam Soon.
• Other Authors: Tan Ching Wai
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer applications::Life and medical sciences
• Online Access: http://hdl.handle.net/10356/19132
• Institution: Nanyang Technological University

In structural bioinformatics, the prediction of the 3D tertiary structure of a protein is a practical problem, which has been attempted by various research groups all around the world. In this project, a predictor will be created to predict long range contacts in 3D protein structures. This is done by first analyzing the structural information of the protein, based on the distances, and creating a predictor using neural networks. Evolutionary profiles will be added to the predictor to analyze how it affects the precision and sensitivity of the results. Results from the neural networks have shown that the predictor obtained better precision and sensitivity with a larger dataset and with evolutionary profile information. Further recommendations to improve on the results of the predictor are to further increase the datasets, provide more input units for neural network training, and to provide the network with more information from the protein.