Replication power in the GIANT consortium for BMI and height.

We tested the novel associations in the UK Biobank discovered by PAT and HIPO for replication in the GIANT consortium. We separately clumped using the lead variant as determined by the m-value. For each variant, we calculate replication power and bin the variants into deciles. The first column lists the trait. The second column is the decile while the third and fourth column are the average power within the set for each respective method. The number of variants tested for replication, the expected number of replications, and the number of variants that replicated are reported in the next six columns. The final two columns contain the number of variants with effect sizes from the GIANT consortium in the same direction seen in the UK Biobank. A binomial test on whether the proportion of effect sizes in the same direction across studies is greater than 50% of all tested variants in the set. A single asterisks means the results are significant at the nominal α = 0.05 and two asterisks indicates significance at .