An integrative algorithm for predicting protein coding regions

Due to the enormous amount of data in DNA sequences to be processed, the computational complexity and speed are important issues to be considered. In this paper, a new integrative method is presented for predicting protein coding regions. We first establish a Takagi-Sugeno fuzzy model to identify the first nucleotide of a codon in coding regions, then the time-frequency characteristics of the output values of the model are analyzed by short-time Fourier transform. The position of coding regions can be exactly identified and the computating time is greatly reduced. Simulation results show the proposed method is more effective for coding regions prediction than the existing coding region discovery tools.