Fig. 5: Performance and computational efficiency comparisons between MultiABEL and GEMMA using simulations. | Heredity

Fig. 5: Performance and computational efficiency comparisons between MultiABEL and GEMMA using simulations.

From: Dual-trait genomic analysis in highly stratified Arabidopsis thaliana populations using genome-wide association summary statistics

Fig. 5

ROC performance of single-trait and dual-trait GWAS results for the target SNP based on genotype data from 199 natural Arabidopsis thaliana inbred lines (Atwell et al.’s dataset) using MultiABEL/GenABEL and GEMMA (a). The effect size of the target SNP was set to \(\beta \propto \sqrt{0.05{h}^{2}}\). The heritability (h2) of each phenotype was set to 0.5, with 10% of genome-wide SNPs randomly selected to simulate population structure. The phenotypic correlation (rP) was set to −0.25. Additional simulation results are presented in Supplementary Figs. S8–15. Computational times for dual-trait GWAS analysis using MultiABEL and GEMMA with the Atwell et al.’s dataset (b) and the 1001 Genomes dataset (c). The Atwell et al.’s dataset comprises 199 individuals with 173,220 SNPs, while the 1001 Genomes dataset includes 1001 individuals with 1,797,898 SNPs. Computation was performed on a single core of the Intel Xeon Platinum 8176 CPU 2.10 GHz processor.

Back to article page