Fig. 1

Workflow for the conditional multi-trait meta-GWAS (CM-GWAS). The CM-GWAS cycles between single-trait GWAS for all traits and a multi-trait meta-GWAS (M-GWAS). After the first cycle, steps from A to D are repeated, jointly fitting the most significant independent M-GWAS variants in a conditional single-trait GWAS for all traits. The t-values from the conditional single trait GWAS (step A) are then updated to re-run the M-GWAS (step B). This continues for one or more cycles until no more significant variants from the M-GWAS are detected in step C. At step C, to determine independence, first the most significant M-GWAS variant from each chromosome is selected and added to the list of putative causal variants. If the pairwise linkage disequilibrium (LD) between this variant and any other significant variant on the same chromosome has r² > 0.1, these other variants are considered as potentially tagging the same causal variant and are not considered as independent QTL for this cycle. Then, from the remaining significant variants in LD r² ≤ 0.1, the next most significant is selected on each chromosome, LD is tested between these and the remaining significant variants and so on, until no more significant variants can be identified in this cycle