Distinguish Coding And Noncoding Sequences In A Complete Genome Using Fourier Transform

Abstract

A Fourier transform method is proposed to distinguish coding and non-coding sequences in a complete genome based on a number sequence representation of the DNA sequence proposed in our previous paper (Zhou et ah, J. Theor. Biol. 2005) and the imperfect periodicity of 3 in protein coding sequences. The three parameters Px(s macr)(1), Px(s macr)(1/3) and Px(s macr)(1/36) in the Fourier transform of the number sequence representation of DNA sequences are selected to form a three-dimensional parameter space. Each DNA sequence is then represented by a point in this space. The points corresponding to coding and non-coding sequences in the complete genome of prokaryotes are seen to be divided into different regions. If the point (Px(s macr)(1), Px(s macr)(1/3), Px(s macr) (1/36)) for a DNA sequence is situated in the region corresponding to coding sequences, the sequence is distinguished as a coding sequence; otherwise, the sequence is classified as a noncoding one. Fisher's discriminant algorithm is used to study the discriminant accuracy. The average discriminant accuracies pc, pnc, qc and qnc of all 51 prokaryotes obtained by the present method reach 81.02%, 92.27%, 80.77% and 92.24% respectively.