Fig. 1

PGS Computation Workflow. Legend: Our genome-wide information metric is simply the direct correlation between full-eur1kg and target-eur1kg PGS, which reflects information loss genome-wide. A correlation of 1.0 indicates that no information loss occurred, whereas a low correlation suggests a substantial loss of information. We derive each ASD-PGS using a C + T method, limiting to biallelic, high quality (info > 0.80) SNPs, for a suite of p-value discovery thresholds (5 × 10^–8, 1 × 10^–6, 10^–4,10^–3, .01, 0.05, 0.10, 0.20, 0.50 and 1.0) scored in PLINK software. Because target-eur1kg-ASD-PGS were clumped and scored for each cohort separately, a different SNP selection informed each of the target-eur1kg PGS, leading to cohort-specific correlations with the full-eur1kg PGS