Parallel Position Weight Matrices Algorithms

Position weight matrices (PWMs) are broadly used in computational biology. The basic problem, Scan, aims to find the occurrences of a given PWM in large sequences. Some other PWM tasks share a common NP-hard subproblem, ScoreDistribution. The existing algorithms rely on the enumeration on a large set of scores or words, and they are mostly not suitable for parallelization.We propose a new algorithm, BucketScoreDistribution, that is both very efficient and suitable for parallelization. We bound the error induced by this algorithm. We realized a GPU prototype for Scan and BucketScoreDistribution with the CUDA libraries, and report for the different problems speedups of 21times and 77times on a Nvidia GTX 280.