Heuristics for pair-wise sequence alignment

Abstract

Develop framework for pairwise alignment of given two sequences based on their feature. There may be situation in which sequences look different but share common structure, function or evolutionary related information. So, rather than forcing alignments to align residue-to-residue comparison, we align them based on segment-to-segment comparison where weight and type is given to each segment. Segments of protein sequence is given based on function, evolution or structure prediction tools. The proposed algorithm is work on proper alignment in higher weighted segments. The rest segments will align with faster existing algorithm so it makes execution faster. The common framework for all type of alignments with one algorithm is new idea. Benchmark Alignment dataBASE , BALIBASE[10] is used to compare result of proposed algorithm with available algorithms. The experiments shows that proposed method does not work for all test cases. But for dissimilar sequences, which looks different, for that proposed method gives good result in less amount of time compare to Needleman-Wunsch[18],Fasta[14] and Blast[24] algorithm.