Extended Data Fig. 8: Computation cost in SAIGE-GENE+ and REGENIE2 by sample sizes for gene-based tests for 18,372 genes with three maximum MAF cutoffs (1%, 0.1%, and 0.01%) and three variant annotations (LoF only, LoF + missense, and LoF + missense + synonymous).
From: SAIGE-GENE+ improves the efficiency and accuracy of set-based rare variant association tests
In total, 165,348 tests were run for each data set. Benchmarking was performed on randomly sub-sampled UK Biobank WES data with White British participants for glaucoma (1,741 cases and 162,408 controls). The reported run times and memory are medians of five runs with samples randomly selected from the full sample set using different sampling seeds. a, Pplots of the time usage and median memory usage in Step 1 as a function of sample size (N). b, Plots of the time usage and median memory usage in Step 2 as a function of sample size (N). Note that singletons only were also included as a mask in the Burden tests in both methods for a fair comparison. SAIGE-GENE+ further automatically output the P values by the Cauchy combination or minimum P-value approach. Tests conducted in the analysis were two-sided.
