Spectral Analysis of Protein Sequences

Show full item record

Title: Spectral Analysis of Protein Sequences
Author: Wang, Zhi
Advisors: Jeffrey Thorne, Committee Member
William Atchley, Committee Co-Chair
Charles Smith, Committee Chair
Bruce Weir, Committee Member
Abstract: The purpose of this research is to elucidate how to apply spectral analysis methods to understand the structure, function and evolution of protein sequences. In the first part of this research, spectral analyses have been applied to the basic- helix-loop-helix (bHLH) family of transcription factors. It is shown that the periodicity of the bHLH variability pattern (entropy profile) conforms to the classical alpha-helix periodicity of 3.6 amino acids per turn. Further, the underlying physiochemical attributes profiles (factor score profiles) are examined and their periodicities also have significant implications of the alpha-helix secondary structure. It is suggested that the entropy profile can be well explained by the five factor score variance components that reflect the polarity/hydrophobicity, secondary structure information, molecular volume, codon composition and electrostatic charge attributes of amino acids. In the second part of this research, complex demodulation (CDM) method is introduced in an attempt to quantify the amplitude of periodic components in protein sequences. Proteins are often considered to be 'multiple domain entities' because they are composed of a number of functionally and structurally distinct domains with potentially independent origins. The analyses of bZIP and bHLH-PAS protein domains found that complex demodulation procedures can provide important insight about functional and structural attributes. It is found that the local amplitude minimums or maximums are associated with the boundary between two structural or functional components. In the third part of this research, the periodicity evaluation of a leucine zipper protein domain with a well-known structure is used to rank 494 published indices summarized in a database (http://www.genome.jp/dbget/aaindex.html). This application allows us to select those amino acid indices that are strongly associated with the protein structure and hereby to promote the protein structure prediction. This procedure can be used to reduce some redundancy of the amino acid indices.
Date: 2005-10-25
Degree: PhD
Discipline: Bioinformatics
URI: http://www.lib.ncsu.edu/resolver/1840.16/5111


Files in this item

Files Size Format View
etd.pdf 1.366Mb PDF View/Open

This item appears in the following Collection(s)

Show full item record