Table 1.

Sample filtering and exclusion criteria from the full dataset

Quality control metric	Individuals excluded
Call rate < 87%	10
Gender discrepancy	1
Sample heterozygosity < 32%	16
Per-family Mendelian errors of >2%	6
Total	33

This table annotates the exclusion of individuals from the initial GWAS. Call rate is based on all SNPs at the outset. The 87% threshold was chosen based on the distribution of missinginess, as was sample heterozygosity. Mendelian errors are impossible configurations of transmission from parent to offspring. A rate higher than 2%, we took to imply either sample mix-up or non-paternity. These exclusions were applied to all individuals in the study.