CAPL: an efficient association software package using family and case-control data and accounting for population stratification

BACKGROUND: With many genome-wide association study (GWAS) datasets available, it is critical that we have statistical tools that are both flexible to accommodate different study designs and fast. We recently proposed the combined APL (CAPL) method, which can use family and case-control datasets and can account for population stratification in the data. Because computationally intensive algorithms are used in CAPL, implementing CAPL with efficient parallel algorithms is essential. RESULTS: We used a hybrid of open message passing interface (open MPI) and POSIX threads to parallelize CAPL, which enable the program to operate in a cluster environment. We used simulations to demonstrate that the parallel implementation of CAPL can analyze a large GWAS dataset in a reasonable time frame when a parallel computing resource is available. CONCLUSIONS: As many GWAS datasets based on both family and case-control designs are available, a flexible and efficient tool such as CAPL will be very helpful to combine the datasets to greatly increase statistical power and finish the analysis in a reasonable time frame.