A Support Vector Machine Based Prediction Model for Lysine Acetylation Site

A Support Vector Machine Based Prediction Model for Lysine Acetylation Site

Chua, Linda

See, Kevin C.

A Research Paper submitted to

Mr. Neil Bretaña

In Partial Fulfillment of the Requirements

for the course in Bioinformatics

July 11, 2012

Abstract

This proposal paper aims to plan the development of a support vector machine based predictor model for lysine acetylation sites in proteins. First, a research on post-translational modification (PTM) in proteins was conducted. Several published related literatures were read and cited. A published paper was chosen as the basis for the entirety of the research. The paper “Predicting post-translational lysine acetylation using support vector machines” by Gnad et. al. was chosen due to its quality and relevance to the research objectives. The dataset was extracted from another source, namely N-Ace (which in turn used the dbPTM database as their source). Six protein features were selected to be the basis for the predictor: (1) frequency, (2) presence, (3) solvent accessibility, (4) hydrophobicity, (5) composition, and (6) BLOSUM63. The datasets extracted were processed. Finally, to attain a set of preliminary results, Weka, a popular data mining and machine learning tool, was used to run the dataset through a support vector machine (SVM).

Introduction

Background of the Study

Proteins are present in every living organism. They are responsible for the many functions of a living organism such as digestion, production of anti-bodies, enzymes, movement, eyesight and more (Bio News Online). But before they become as such, they undergo a process known as protein biosynthesis which occurs inside cells. In a nutshell, this process takes 20 amino acids and forms them into a long chain called polypeptides (Walker, 2012). After the formation of this long chain, a protein is made. But before it can become fully functional, it undergoes another process known as post-translational modification or abbreviated as PTM....