A linear space algorithm for computing maximal common subsequences
Communications of the ACM
Parallel Biological Sequence Comparison Using Prefix Computations
IPPS '99/SPDP '99 Proceedings of the 13th International Symposium on Parallel Processing and the 10th Symposium on Parallel and Distributed Processing
Hi-index | 0.00 |
Aligning long DNA sequences is a fundamental and common task in molecular biology. Though dynamic programming algorithms have been developed to solve this problem, the space and time required by these algorithms are still a challenge. In this paper we present the Parallel Linear Space Alignment (PLSA) algorithm to compute the long sequence alignment to meet this challenge. Using this algorithm, the local start points and grid cache partition the whole sequence alignment problem into several smaller independent subproblems. A novel dynamic load balancing approach then efficiently solves these subproblems in parallel, which provides more parallelism in the trace-back phase. Furthermore, PLSA helps to find k near-optimal non-intersecting alignments. Our experiments show that this proposed algorithm scales well with the increasing number of processors, and it exhibits almost linear speedup for large-scale sequences.