SOAP3

SOAP3 accelerates alignment of short DNA sequence reads to a reference genome by leveraging graphic processing units (GPUs) for high-throughput, exhaustive short-read mapping.

Key Features:

• GPU Acceleration: Utilizes GPU multi-processors to parallelize read alignment and increase throughput.
• BWT-based Indexing: Builds upon the compressed full-text index (BWT) framework from SOAP2 for sequence indexing and alignment.
• Speed Performance: In benchmarks with Illumina HiSeq 2000 100 bp reads, aligned one million read pairs to the human reference genome in less than 30 seconds, outperforming BWA by at least 7.5× and Bowtie by at least 20×.
• Comprehensive Alignment: Non-heuristically reports all possible alignments for reads with up to four mismatches, yielding slightly higher alignment rates than BWA and Bowtie.

Scientific Applications:

• Whole-genome sequencing: Rapid alignment of short reads to reference genomes for WGS projects.
• Variant calling: Generation of comprehensive alignments as input for variant discovery workflows.
• High-throughput sequencing data processing: Processing large Illumina HiSeq datasets where speed and exhaustive mismatch handling are required.

Methodology:

Builds on the compressed full-text index (BWT) framework from SOAP2, implements GPU multi-processor parallelization, and performs non-heuristic exhaustive alignment reporting for reads with up to four mismatches.