Fig. 3: Each point represents the mean value of 10 replications, while the error bars are represented by the estimate ± 1.96 standard errors.

Run times were assessed for a varying number of individuals and SNPs. The number of SNPs (M) varies from 100k to 1 million, while the number of individuals vary from 10k to 100k. SPACox uses a single CPU core, as no parallelization is available. We used 1, 2, and 4 CPU cores for estimating the ADuLT phenotype and performing the linear regression GWAS for this phenotype. The means and corresponding standard errors of the run times can be found in Supplementary Data 1.