Parallelization of RNA secondary structure prediction algorithm on multicore

Abstract

RNA secondary structure prediction is invaluable in creating new drugs and understanding genetic diseases, in gene expression and regulation. RNA molecules have proven their clinically consequential deep inroads into the biological functions. Needless to mention about the role of even small RNA's with numerous nucleotides which does the function of gene splicing, editing, and regulation. There is accelerating increase of data set sizes originating from powerful high-throughput measuring devices in Bioinformatics, A major limiting factor in achieving good performance on modern architectures is memory latency and bandwidth. The cost of computation is very less than the cost of moving data from main memory. This disparity between communication and computation costs has compelled the programmers to make changes in the ways algorithms are designed.A promising solution to speed up sequential programs, which are difficult to parallelise otherwise is through Speculative parallelisation.This paper proposes a new software-only speculative parallelization scheme for implementing RNA Secondary Structure Prediction algorithm in parallel. The scheme is developed after a systematic evaluation of the design options available. It is also shown to be efficient, robust and to outperform previously proposed schemes used for parallel implementation of RNA Secondary Structure Prediction.