In the MCTFR, ancestry was based initially on the ethnicity specified on a birth certificate, adoption records, or by self-report. Of the 8405 GWAS samples, 7599 (90.4%) self-reported as having primarily European ancestry (i.e., “white”), 382 (4.5%) as Asian, 83 (1%) as African-American (i.e., “black”), and 127 (1.5%) reported mixed ancestry. All other ethnicities had a self-reported frequency of < 1% and there were approximately 1% with a missing self-report. Because the genetics of complex phenotypes can vary across different ancestral groups (Bamshad, 2005), the primary sample used in our GWAS analysis will be comprised of individuals of European ancestry. However, to improve on the accuracy of the self-report data and to deal with missing self-report data, we ran EIGENSTRAT (Price et al., 2006) analysis, extracting the first 10 principal components (PCs), to aid in the identification of a cluster of individuals with European ancestry.
