Logistic regression as implemented in PLINK is a powerful and commonly used framework for assessing gene-gene (GxG) interactions. However, fitting regression models for each pair of markers in a genome-wide dataset is a computationally intensive task. Performing billions of tests with PLINK takes days if not weeks, for which reason pre-filtering techniques and fast epistasis screenings are applied to reduce the computational burden.
Here, we demonstrate that employing a combination of a Xilinx UltraScale KU115 FPGA with an Nvidia Tesla P100 GPU leads to runtimes of only minutes for logistic regression GxG tests on a genome-wide scale. In particular, a dataset of 53,000 samples genotyped at 130,000 SNPs was analyzed in 8 min, resulting in a speedup of more than 1,000 when compared to PLINK v1.9 using 32 threads on a server-grade computing platform. Furthermore, on-the-fly calculation of test statistics, p-values and LD-scores in double-precision make commonly used pre-filtering strategies obsolete.
Genome-wide association study (GWAS) Genome-wide interaction study (GWIS) Gene-gene (GxG) interaction Linkage disequilibrium (LD) BOOST Hardware accelerator Hybrid computing Heterogeneous architecture
This is a preview of subscription content, log in to check access.
Atkinson, M.D., Sack, J.R., Santori, N., et al.: Min-max heaps and generalized priority queues. Commun. ACM 29(10), 996–1000 (1986)CrossRefGoogle Scholar
Cattaert, T., Calle, M.L., Dudek, S.M., et al.: Model-based multifactor dimensionality reduction for detecting epistasis in case-control data in the presence of noise. Ann. Hum. Genet. 75(1), 78–89 (2011)CrossRefGoogle Scholar
Genetic Analysis of Psoriasis Consortium, et al.: A genome-wide association study identifies new psoriasis susceptibility loci and an interaction between HLA-C and ERAP1. Nat. Genet. 42, 985–990 (2010). https://doi.org/10.1038/ng.694
Goudey, B., Rawlinson, D., Wang, Q., et al.: GWIS: model-free, fast and exhaustive search for epistatic interactions in case-control GWAS. Lorne Genome 2013 (2013)Google Scholar
Hu, X., Liu, Q., Zhang, Z., et al.: SHEsisEpi, a GPU-enhanced genome-wide SNP-SNP interaction scanning algorithm, efficiently reveals the risk genetic epistasis in bipolar disorder. Cell Res. 20, 854–857 (2010)CrossRefGoogle Scholar
Ibrahim, Z.M., Newhouse, S., Dobson, R.: Detecting epistasis in the presence of linkage disequilibrium: a focused comparison. In: 2013 IEEE Symposium on CIBCB, pp. 96–103 (2013). https://doi.org/10.1109/CIBCB.2013.6595394
Kässens, J.C., Wienbrandt, L., et al.: Combining GPU and FPGA technology for efficient exhaustive interaction analysis in GWAS. In: 2016 IEEE 27th International Conference on ASAP, pp. 170–175 (2016). https://doi.org/10.1109/ASAP.2016.7760788
Kässens, J.C.: A hybrid-parallel architecture for applications in bioinformatics. No. 2017/4 in Kiel Computer Science Series, Department of Computer Science, CAU Kiel (2017). Dissertation, Faculty of Engineering, Kiel University. https://doi.org/10.21941/kcss/2017/4
Kirino, Y., Bertsias, G., Ishigatsubo, Y., et al.: Genome-wide association analysis identifies new susceptibility loci for Behçet’s disease and epistasis between HLA-B*51 and ERAP1. Nat. Genet. 45, 202–207 (2013). https://doi.org/10.1038/ng.2520CrossRefGoogle Scholar
Lee, S., Kwon, M.S., Park, T.: CARAT-GxG: CUDA-accelerated regression analysis toolkit for large-scale gene-gene interaction with GPU computing system. Cancer Inform. 13s7, CIN.S16349 (2014). https://doi.org/10.4137/CIN.S16349
The Australo-Anglo-American Spondyloarthritis Consortium (TASC), et al.: Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility. Nat. Genet. 43, 761–767 (2011). https://doi.org/10.1038/ng.873