Efficient Exonic Regions Prediction in DNA Sequence Using Fast Converged Adaptive Filter

Abstract

Signal processing takes an important role in genomic with enormous data available in public domain. Generally digital filters are applied to predict the protein and genes, but it needs to be redesigned when the characteristic frequency and periodic behavior is changed. In this paper proposed the novel adaptive algorithm which can identify the genes and proteins effectively from unified framework. First using the electron ion potential method the symbolic DNA sequences are converted in to digital signal. Secondly the filtering scheme for genomic signal processing with periodic behavior in biological sequence is introduced, which can predict and analyze the biological region that are interested in. finally the proposed adaptive filtering method is applied to recognize the exons of protein coding regions according to periodic-3 property. The exons prediction curves are obtained with Discrete Fourier Transform (DFT), Least mean square (LMS), and proposed Fast Recursive least Mean Square (F-RLS) algorithms. It is shown that proposed method shows efficiency in convergence of identification and precise prediction of exons regions compared to existed methods.