Following uniform quality control and imputation using the 1000 Genomes Project (phase 3) (17) in the anorexia nervosa case-control cohorts, we performed association analysis using an additive model using the dosage data for each cohort. Following adjustment for unbalanced case and control numbers across our 12 strata [see (18)], our summed effective balanced sample size was 5082 cases and 5082 controls. Accordingly, our power was 83.1% for a genotype relative risk of 1.25, at an allele frequency of 0.2 at p<5×10−8 (http://zzz.bwh.harvard.edu/gpc). Analysis within datasets was performed in PLINK with the first ten principal components as covariates. METAL (18) was used to conduct fixed-effects meta-analysis across the twelve datasets using inverse-variance weighting. Results were obtained for 10,641,224 single nucleotide polymorphisms (SNPs) and insertion-deletion variants with minor allele frequency > 1% and imputation quality scores > 0.6 [see Supplementary Figure S2 for quantile-quantile (QQ) plot]. The GWAS statistic inflation (λ) was 1.080 with a sample size adjusted λ1000 of 1.008, consistent with minimal population stratification or other systematic biases. Plotting was performed in R (19) (see Supplementary Text for additional methods and quality control details and Supplementary Table S1 for individual study details).
