Performance analysis of different DNA to numerical mapping techniques for identification of protein coding regions using tapered window based short-time discrete Fourier transform

Abstract

Prior to applying the digital signal processing techniques for identification of protein coding regions, mapping of DNA alphabet into numerical sequences is necessary. In this paper, the performance of existing DNA to numerical mapping techniques is analyzed at the nucleotide level for the identification of protein coding regions using tapered window based short-time discrete Fourier transform (ST-DFT) method applied to the GENSCAN test set. The results show better identification accuracy for Voss, z-curve and tetrahedron mapping technique as compared to other mapping techniques when tapered windows are used for ST-DFT method.