Commonly used genomic arrays may lose information due to imperfect coverage of discovered variants for autism spectrum disorder

Table 2 Sample size, variant numbers, and coverage statistics for top ASD hits per cohort

Cohort (Array)	N measured Variants surviving QC	N Variants surviving post-imputation QC	N Top ASD Variants on GWA Array^a	N Top ASD Variants surviving post-imputation QC	N SNPs in Score^b	ASD-PGS coverage metric^c
IBIS (5 M)	2,400,509	38,343,801	52 (31)/88	88/88	249,698	0.9569
EARLI (5 M)	2,525,262	38,123,095	54 (32)/88	88/88	242,199	0.9445
MARBLES (1 M)	578,578	33,317,727	28 (11)/88	82/88	229,119	0.9295
SEED (GSA)	877,115	32,275,019	31 (11)/88	83/88	268,035	0.9849

^aCoverage reflects the presence on the panel of either the discovery variants itself or at least one proxy SNP. Number in parentheses refers to the non-proxy number
^bFull-eur1kg containes 281,593 SNPs after clumping
^cCoverage metric reflects the correlation of full-eur1kg and target-eur1kg scores as explained in Fig. 1

ISSN: 1866-1955