FastMSECT Algorithm: A Many-Core Fast Scalable Method for Massive String Exhaustive Comparison Technology such as GMSECT

BIRAC E-Yuva Centre, Bhopal, Madhya Pradesh, India

^* Corresponding author: abi@abitoeq.net

Abstract

FastMSECT is a divide-and-conquer based algorithm to deal with large strings such as DNA sequences of Genome size in many-core processors or co-processors. FastMSECT uses the power of many core processors to adequately match the L3 cache sizes which is uniformly available to all the CPU cores as shared memory and in doing so it expedites the process of string comparison. Sequence comparison algorithms such as Smith Waterman usually have quadratic or exponential time complexity, which FastMSECT drops the time complexity to linear growth with longer sequence lengths. FastMSECT algorithm is implemented in the form of a tool called GMSECT for Genome-wide Massive Sequence Exhaustive Comparison Tool, which is made available via web-service of GenomeKlinik. GenomeKlinik is a versatile internet-based application where biomedical scientists and clinical practitioners can feed in their genomic sequence of interest and retrieve genomic variation as compared to a reference genome such as HuRef for human genome, and also get downstream annotation for the variation such as disease phenotype and features. It internally makes use of GMSECT, SQL and other commercially freely available software tools. The tool can also be used for large scale population data study and novel variation detection, and can be downstream channeled into biomarker discovery. GenomeKlinik although focuses on genomic sequences such as the Next Generation Sequencing (NGS), it is also useful for other kinds of data such as epigenome and other omics. GenomeKlinik serves as a one stop bioinformatics platform where the customer can create recommended pipeline for data processing and get meaningful information such as for the patient whose Exome or Genome or Epigenome data has been obtained. The web application server is a Freemium mode of offering limited services for free.