The Search for Gene-Gene Interactions in Colorectal Cancer: Using HPC to Overcome Computational Barriers

UK National Cancer Registration data indicates that some 35,000 people each year are diagnosed with colorectal cancer (cancer of the large bowel and rectum) and 16, 000 die from the disease. The Colon Cancer Genetics Group (CCGG) at the University of Edinburgh investigates the relationship between genetic markers and colorectal cancer by using a significant part (560,000 markers, 1000 cases, 1000 controls) of the biggest genotypic data set for large bowel cancer. However, the analysis is virtually intractable for a PC-based researcher (theoretical runtime of 400 days; 3.3TB of memory and hard disk space). CCGG collaborated with EPCC, the super computing centre of the University of Edinburgh, to optimise and parallelise the analysis code. We achieved a runtime of approximately 5 hours on 512 processor cores on HECToR, the national supercomputer of the UK. The use of EPCC's skills and HPC resources has enabled CCGG to explore new territory for genetic marker analysis in colorectal cancer.