Notably all of the top menarche variants had MAF ≥7%. While it has been suggested that low-frequency variants have larger effects than common variants39, we were clearly underpowered to detect low-frequency variants (MAF <5%) with modest effect sizes. It is also possible that rare variants are not well captured using genome-wide chips. Future re-imputation using deep sequencing data from the 1000 Genomes Project may identify additional low frequency hits as well as refining the location of possible functional variants.
