A Data Parallel Strategy for Haplotype Assembly Problem on Multi-core Computers

Haplotype assembly problem is one of the core problems in the whole genome research, and a lot of methods have been proposed to solve this problem. However the computation time will increase greatly in the case of large-scale data for most methods. In this work, we propose a parallel strategy Par HA (Parallel Haplotype Assembly) for haplotype assembly problem based on data parallelism on multi-core computers. The SNP data were divided into several data blocks, which were processed in parallel, and the final haplotypes were constructed by fusing the results of data blocks. We have tested Par HA in terms of the reconstruction rate and the speedup, and extensive experimental results indicate that the proposed method will improve the computation efficiency significantly while achieving comparative reconstruction rate.