high performance computing on graphics processing units: hgpu.org

Exact short read mapping to whole genomes with the Smith-Waterman algorithm is computationally expensive yet highly accurate when aligning reads with mismatches and gaps. We introduce a GPU program called MaxSSmap with the aim of achieving comparable accuracy to Smith-Waterman but with faster runtimes. Similar to mainstream approaches MaxSSmap identifies a local region of the genome followed by exact alignment. Instead of using hash tables or Burrows-Wheeler in the first part, MaxSSmap calculates maximum scoring subsequence score between the read and disjoint fragments of the genome in parallel on a GPU and selects the highest scoring fragment for exact alignment. We evaluate MaxSSmap’s accuracy and runtime when mapping simulated E.coli and human reads of 10% to 30% mismatches with gaps of various lengths to the E.coli genome and human chromosome one respectively. We show that MaxSSmap attains comparable high accuracy and low error to fast Smith-Waterman programs yet has much lower runtimes. We also show that MaxSSmap can map reads rejected by fast mappers with high accuracy and low error much faster than if Smith-Waterman were used. The MaxSSmap source code is freely available.