Each study performed genome-wide association testing for age at menarche across approximately 2.5 million SNPs, based on linear regression under an additive genetic model. Analyses were adjusted for birth year in order to remove the effect of the temporal decline in age at menarche. Studies used PLINK, ProABEL, MACH2QTL, SNPTEST, R-packages or MERLIN-fastassoc for the association testing. The results from individual studies were corrected by their respective genomic inflation factors (lambda) (Supplementary Table 1) according to the genomic control (GC) method to correct for population stratification38.
