A tri-nucleotide mapping scheme based on residual volume of amino acids for short length exon prediction using sliding window DFT method

One of the great challenges in the field of bioinformatics is how to locate the accurate protein-coding regions in a given DNA sequence. The accurate identification of the protein-coding region is useful in many applications. For instance; it helps in characterizing new proteins, drug designing, and also in revealing the evolutionary background of a particular organism. DSP based techniques are quite popular in the protein-coding region identification. The first essential step of the DSP based exon prediction technique is to convert the base sequences into the numerical sequence. The choice of the numerical mapping scheme affects how well the characteristic feature of the DNA sequence is reflected in the numerical domain which helps in finding the accurate location of exons. In the last two decades, numbers of mapping schemes have been successfully used for exon prediction. However, locating the short length exon is still a difficult task. In this paper, we have proposed a tri-nucleotide mapping scheme that exploits the residual volume property of amino acid to encode the given DNA sequence. It is obtained that the proposed tri-nucleotide mapping scheme provides better results than other mapping schemes in case of short length exon detection.