Computational Methods to Locate and Reconstruct Genes for Complexity Reduction in Comparative Genomics

Abstract

Discovering the functions of proteins in living organisms is an important tool for understanding cellular processes. The source data for such analysis are commonly the peptide sequences. Most common algorithms used to compare a pair of nucleotide sequence are Global alignment algorithm (Needleman-Wunch algorithm) or local alignment algorithm (Smith-Waterman algorithm). Analysis of these algorithms show that time complexity required to the above mentioned algorithms is O(mn) and space complexity required is O(mn), where m is size of one sequence and n is size of the other sequence. This is one of the major bottlenecks as most of the sequences are very large. The proposed Coding Region Sequence Analysis(CRSA) algorithm presents a method to reduce both time and space complexity by meaningfully reducing the size of sequences by removing not so significant exons using wavelet transforms. DSP techniques supply a strong basis for regions identification with three-base periodicity.