epiSNP_cc: A Cloud Computing GWAS Software for Epistasis Testing

Dense single nucleotide polymorphism (SNP) markers covering the entire genome provide a powerful molecular tool for understanding genome-phenotpye relationships, but pairwise epistasis testing of SNP markers is computationally demanding due to the large number of pairs. In this research, we developed epiSNP_cc suite of software tools for pairwise epistasis testing based on our previous supercomputer parallel computing tools for genome-wide association analysis. With epiSNP_cc, users without access to a supercomputer or a computer cluster may conduct large scale epistasis tests using commodity processors available via cloud computing services. An image of epiSNP_cc is available on Amazon EC2 and accessible to the public for free. The epiSNP_cc image can be accessed at console.aws.amazon.com by searching through the community Amazon Machine Image (AMI). Running on standard EC2 hardware instances makes it easier for users to execute their own analysis with epiSNP_cc. Once the image is selected and deployed, for each epiSNP_cc software tool, a binary executable is available with an example of how to run it. The user does not need to compile or install anything. No software license is required either. Once a job is finished running, the user can terminate the instance and save work for future runs. In addition to running in the cloud, the software can run from a single portal to a wide variety of compute resources ranging from a desktop to large computer clusters including supercomputers at national centers such as XSEDE.