SPRINT training course: user-friendly statistical analysis using HECToR, the UK’s most powerful supercomputer
SPRINT (Simple Parallel R INTerface) provides easy access to high performance computing (HPC) for the analysis of high-throughput, post-genomic data using the statistical programming language R. The first ever SPRINT training course will be run at NAG Oxford on December 1st 2011. The course aims to give a general overview of the hardware, software and configuration of the HECToR system in the context of using SPRINT.
The first full version of the SPRINT framework - SPRINT v1.0.0 - will be released this month. SPRINT gives R users the parallel tool kit they need to perform computationally-challenging statistical analyses easily on HPC. It lifts the limitations on data size and analysis time commonly found in many disciplines such as bioinformatics, where large volumes of highly parallel, high throughput post-genomic data are generated. SPRINT is unique in its HPC platform independence, its focus on complex, hard to parallelise, data-dependent functions and its user friendliness.
SPRINT contains optimised versions of the following R functionality: pair-wise Pearson’s correlation, permutation testing, random forest, rank product, partitioning around medoids and apply. It is available to use on HECToR, the UK’s national academic supercomputing service, which is hosted by EPCC.
SPRINT is an open-source project and is open to contributions. All sources are available from R-Forge