DNA Sequence Analysis Based on Fuzzy Neural Network and Simulated Annealing Technique

• Main Authors: Shaw-Hwa Su, 蘇紹華
• Other Authors: Hong-Yuan Liao
• Format: Others
• Language: en_US
• Published: 1994
• Online Access: http://ndltd.ncl.edu.tw/handle/01804767199317031064

碩士 === 國立中正大學 === 電機工程研究所 === 82 === Protein plays an important role in all kinds of animals. The whole procedure of protein synthesis includes conversions of DNA sequence to primary RNA, primary RNA sequence to mRNA, and finally mRNA sequence to protein. Usually, the nucleotide acids in a DNA sequence can be labelled into two different ca- tegories, i.e., the coding nucleotide acid and the noncoding nucleotide acid. Usually, in the protein synthesis process the noncoding regions in a DNA sequence have to be separated from the coding regions. Therefore, how to identify and remove the noncoding nucleotide acids in an unknown DNA sequence be- comes an important issue in the protein synthesis process. In the first part of this thesis, we propose a fuzzy neural network to deal with the above problem. The proposed approach takes the neighborhood information of a given nucleotide acid into account and the experimental results reflect that the method is superb. After all the DNA sites in a sequence have been identified, the next problem to be solved is to perform sequence alignment. In this thesis, a method which can perform simultaneous alignment on a large amount of sequences is presented. In this method, the simulated annealing technique is adopted to execute an opti- mization process. From the identified consensus sequence plus the calculation if the Shannon information clearly reflect that several sequence motives are conserved:(a) a guanosine rich re- gion, (b) the polypyrimidine tract.