Frequency-based re-sequencing tool for short reads on graphics processing units

Abstract

After the reference genomes of many organisms are sequenced in the post-genetic era, an important issue is to do the re-sequencing of individual genomes with high-throughput reads. Many next-generation sequencing machines have been proposed in the last few years and a series of re-sequencing tools have been developed for mapping short reads to the reference genome. FRESCO is a frequency-based re-sequencing tool without using hash look-up table algorithm and Burrows-Wheeler transformation. FRESCO offers more flexibility in the mapping and then obtains satisfactory mapping results. However, FRESCO is a computation-intensive tool. Therefore, in this paper, a tool, CUDA-FRESCO v1.0, was proposed to reduce the computation time of FRESCO by using the graphics processing units with CUDA. By comparing to FRESCO, CUDA-FRESCO v1.0 achieved 63× speedups for the mapping kernel and 20× speedups for the overall computation time.