Publication: Prediction of antimicrobial peptides using a complexity-based distance measure
Loading...
Date
2012-06-01
Authors
Loo, Yue Lin
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This project is concentrated on the attempt to establish a new classification method for predicting Antimicrobial Peptides (AMPs). AMPs represents one of the classes in the peptides that act as a part of immune system. Since the experiment to extract AMPs from protein sequences are costly and require longer time to setup, a prediction tool based on the computational method needs to be developed to solve the problems. One of the computational methods, which is known as “complexity distance method” is used to develop the prediction tools for AMPs. Lempel-Ziv (LZ) complexity algorithm and Nearest Neighbour algorithm are implemented into this project to predict the AMPs. After completing the prediction tool, the performance on predicting AMPs will be compared to other recentlydeveloped methods, such as sequences alignment and feature selection. Complexity distance measure method shows the best result compared to feature selection method based on sensitivity. The sensitivity of complexity distance measure is 82.05% and sensitivity of sequence alignment is 90.87%. Both methods used the same 150 test sequences from test database and 12766 training sequences. However,
sequence alignment method is shown to have a better performance compared to complexity distance measure. The sensitivity of complexity distance measure is 48.67 % and sensitivity of feature method is 46.00%. Both methods used the same 986 test sequences from test database and 12766 training sequences. For future improvement,
integration between sequence alignment method and complexity distance measure may
be developed