关键词: DNA序列编码区, 密码子, Takagi-Sugeno模糊模型, 模糊聚类, 最小二乘法

Abstract: An important step in gene identification is to predict coding regions in DNA sequence.Due to the large volume of gene data leading to the problem of poor generalization capability and lower computing speed in many algorithms of prediction of coding region.In this paper，a Takagi-Sugeno model of DNA sequence is built based on the different composition of nucleotides in coding regions and non-coding regions.First，the system is quickly divided into several fuzzy parts using clustering algorithm based on the fuzzy likelihood function．Then，regarding cluster number as a rule number，Takagi—Sugeno fuzzy model has been built．Finally，the consequent parameters of the model are identified associating with LS.The algorithm not only can predict coding regions，but also can identify the first nucleotide of the codon in coding regions.This is very significant for accurate translation into a protein sequence.The algorithm is simple and simulation results show the proposed method is more effective for coding regions prediction than the existing coding region discovery tools.

Key words: coding region in DNA sequence, codon, Takagi-Sugeno model, clustering algorithm, Least Square（LS）