Dissertation > Excellent graduate degree dissertation topics show

Study on Coding Schemes of Protein Secondary Structure Prediction Based on Support Vector Machines

Author: YeXiaoJiao
Tutor: LiWangGen
School: Anhui Normal University
Course: Applied Computer Technology
Keywords: Encoding method Protein Support Vector Machine Secondary Structure Prediction
CLC: Q51
Type: Master's thesis
Year: 2011
Downloads: 35
Quote: 0
Read: Download Dissertation

Abstract

Significant protein structure research, analysis of protein structure, function and relationship is an important part of the post-genome project. However complex and time-consuming with traditional experimental determination methods, theoretical analysis method to predict protein structure has become increasingly important. Direct prediction of protein structure more difficult, relatively simple protein secondary structure prediction became predict the spatial structure of the bridge from the primary sequence of the protein. Protein structure prediction methods and coding methods are important factors affect prediction accuracy rate, this paper studies encoding method for protein secondary structure prediction, and a comprehensive single sequence protein secondary structure prediction problem encoding method. First, this paper introduces the research background and significance of the secondary structure prediction, and introduces some basic background knowledge, such as knowledge of bioinformatics, molecular composition and structure of the protein, protein structure prediction, especially secondary structure prediction a common method, and its prediction result evaluation method. Shortly thereafter, the article describes a support vector machine. Because the protein secondary structure prediction is actually a pattern recognition problems, support vector machine method exhibit many unique advantages in solving small sample size, nonlinear and high dimensional pattern recognition, this paper uses the method to predict the secondary structure. Then study support vector machines in high-dimensional pattern recognition problem, their personal credit rating as instances of research and get a better prediction results. Finally, several amino acids encoding a comparative study and analyze their respective characteristics and defects. For a single sequence of non-homologous or low homologous protein secondary structure prediction problem, a new coding method. The encoding is based on the tendency of amino acids appear in the secondary structure of each factor, and the amino acid hydrophobicity value is classified, and is represented in binary form to each class of amino acids. Modeling was carried out under the same experimental conditions prediction after a new encoding method and other encoding methods, experimental results show that the new coding method to more fully utilize the protein primary structure information, more suitable for the non-homologous or homologous protein structures forecast.

Study on Coding Schemes of Protein Secondary Structure Prediction Based on Support Vector Machines

Abstract

Related Dissertations