For the NTR cohort, the statistical power to detect a SNP at the genome-wide significance level that explains 0.1 % of the true phenotypic variance (latent trait) with an allele frequency of 0.5 when using only the 12 NEO-FFI items was 18 % (N = 5,299 individuals with NEO-FFI data on Neuroticism) and increased to 44 % when using IRT scores based on both NEO-FFI and ABV data (N = 31,309 individuals with either NEO-FFI data, ABV data or both). In the QIMR-adult sample, the power with only 12 NEO-FFI items was 0 % (N = 3,712). Using all available data from all inventories and analyzing IRT scores yielded a power of 30 % (N = 26,692). Thus, the power in GWAS substantially increases if item data from multiple inventories are included, if available.
