Analysis of Genomic and Proteomic Sequences Using Weight Functions and Discrete Transforms through Simulation Studies

Abstract

At present Signal Processing is in the midst of a major transition froma focus on classical newlinesignals in Electrical Engineering applications to a much wider usage devoted to the newlineanalysis of signals in a broad spectrum of science and engineering disciplines. The tools newlineof Signal Processing are currently being employed to embark on new frontiers in science newlineand technology by analysis of signals in diverse fields of astronomy, energy, finances, newlinegenomics, geosciences, privacy, security, social networks, and much more. newlineThis work explores and establishes the applicability of Signal Processing newlinealgorithms in the revelation of bio-molecular systems. Focus is on the development of newlinealgorithms for the analysis of Deoxyribonucleic acid (DNA) sequences and identification newlineof protein interaction sites using transforms, weight functions and digital filters. DNA newlinesequences have been analyzed to discriminate protein coding regions from the noncoding newlineones. Amino acid sequences have been subjected to analysis to identify newlineinteraction sites of a protein with other molecules. These interaction sites are referred to newlineas hot spots and decide the function of protein molecules. newlineApart from Signal Processing based methods statistical methods based on newlinepreviously known database, that is used to train a supervised classifier like Markov newlinechains, have also been applied for the analysis of genomic sequences. However, Signal newlineProcessing algorithms are model independent in contrast to statistical methods that are newlinemodel dependent. The training dependence reduces the adaptability of statistical methods newlinefor new sequences from unknown organisms with no or small training sets. Likewise, newlineDSP based algorithms can also be used for analyzing and identifying hot spots in newly newlinediscovered proteins as they do not require the structural information of proteins. newlineExperimental methods on the other hand are expensive, time consuming and require a lot newlineof efforts. So wet lab experiments can be selectively performed by using the estimates newlineobtained from DSP based computational