A fast heuristic algorithm for similarity search in large DNA databases

The similarity search is an important procedure in genomic research. Considering enormous DNA sequence databases, it becomes impractical to use basic pattern matching methods such as the Smith-Waterman algorithm to compute similarities between two sequences. In this paper, we propose an efficient query processing method based on indexing. It uses little storage and rapidly finds the similarity between two sequences in the DNA sequence database. At first, our algorithm partitions sequences into equal length windows. And then it transforms subsequences in each window into a multidimensional vector space by indexing frequencies of characters, which includes the positional information of characters in subsequences, as weight. Our algorithm not only has linear time complexity but enhances the accuracy of query processing. The result of experiments shows that our algorithm performs query processing more accurately and has a lower error ratio than the well-known BLAST and heuristic algorithms which use only frequencies of characters.