Implementation of a novel algorithm for pairwise protein alignment and the analysis of clustering of dihedral angles

With more and more complete genome sequencing being discovered, structural genomics projects are formed with the aim to determine 3D structure of all proteins. This has led to an increasing number of solved protein 3D structures in recent years. Hence, this has raised the need to develop an efficien...

Full description

Saved in:
Bibliographic Details
Main Author: Nesy.
Other Authors: Tan Ching Wai
Format: Final Year Project
Language:English
Published: 2009
Subjects:
Online Access:http://hdl.handle.net/10356/16987
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:With more and more complete genome sequencing being discovered, structural genomics projects are formed with the aim to determine 3D structure of all proteins. This has led to an increasing number of solved protein 3D structures in recent years. Hence, this has raised the need to develop an efficient structural alignment algorithm to analyze the similarity of protein 3D structures and predict the unknown protein functions. In this project, a pairwise protein alignment was developed to compare the similarity of two protein structures by using the dihedral angles calculated from their 3D coordinates. In later part, clustering was performed on the dihedral angles in order to cluster similar fragments of protein structure together. Due to a change in research direction instructed by supervisor, pairwise alignment results will not be shown in this report. Clustering was performed on subsequences of 781 dihedral angle files, where similar fragments of dihedral angles were clustered into different number of clusters. Dihedral angles were manipulated into window sizes of 1, 3, 5, 7, 9, and 11 and clustered into 2 to 10 clusters. Cluster results were further processed with more information and analyzed. The cluster results showed that clustering of dihedral angle fragments with window size 9 and cluster numbers of 4 appeared to be most informative. It also successfully divided the fragments according to the different type of residues. The results also showed that window size 9 is good for protein fragment, as a similar clustering approach done on protein sequences also yielded similar results [10]. Clustering of dihedral angle fragments can be seen as the first step to move towards the computational analysis of protein structure. For future works, the cluster analysis results can be extended for further understanding in similar fragments of proteinstructure and be used in structural alignment.