A parallel hash-based method for local sequence alignment

Algorithms utilize an index-based aligning strategy, like a hash table, which typically entails the seed-and-extend method and is a time-consuming task. Here, we developed a hash-based search algorithm based on the SSAHA method without the use of seed-and-extend to conduct search and alignment faster than previous methods with multiple processors. In the proposed method by using the overlapping method in query and reference sequences, the accuracy and sensitivity increased. Further, the speed also increased by creating a hash table for the reference sequence when it was placed in the memory. Furthermore, by evaluating three datasets of different sequences in size and volumes, the effect of the created piece lengths as well as the effect of multiple processors on each dataset was evaluated indicating not only appeasing the time issue in alignment but also improving the mapping speed compared to the BLAST and SSAHA algorithms.