TABLE 2.

Summary information of the GWAS datasets employed in our sex-specific two-sample MR analysis.

Trait	Sample size (case/control)	k0	M	References
BMI
Sex-combined	322,154	97	2,517,828	Locke et al., 2015
Female-specific	171,977	38	2,459,695	Locke et al., 2015
Male-specific	152,893	30	2,443,565	Locke et al., 2015
WHR
Sex-combined	210,086	39	2,542,431	Shungin et al., 2015
Female-specific	116,742	34	2,467,778	Shungin et al., 2015
Male-specific	93,480	3	2,146,136	Shungin et al., 2015
WC
Sex-combined	231,355	70	2,545,772	Shungin et al., 2015
Female-specific	127,470	25	2,473,035	Shungin et al., 2015
Male-specific	104,079	29	2,294,965	Shungin et al., 2015
HIP
Sex-combined	211,117	89	2,540,653	Shungin et al., 2015
Female-specific	117,340	41	2,466,814	Shungin et al., 2015
Male-specific	93,965	31	2,188,855	Shungin et al., 2015
Breast cancer	228,951 (122,977/105,974)
Female-specific			13,011,123	Michailidou et al., 2017
Prostate cancer	140,306 (79,194/61,112)
Male-specific			16,486,833	Schumacher et al., 2018

BMI, body mass index; WHR, waist-to-hip ratio; WC, waist circumference; HIP, hip circumference; SNP, single-nucleotide polymorphism; GWAS, genome-wide association study; MR, Mendelian randomization.

Note that k0 denotes the number of candidate instruments that were obtained directly with those associated SNPs (P < 5.00E-8) reported in the original GWAS papers; M is the total number of SNPs.