550k duo

The Illumina 550K, 550K duo, and 610K arrays were used for genotyping samples with a call rate below 97.5%, gender mismatch, excess autosomal heterozygosity (>0.336), duplicates or family relations and ethnic outliers were excluded. Genetic variants were filtered by Hardy-Weinberg equilibrium (P<10^-6), allele frequency (excluding minor allele frequency (MAF < 0.001) and SNP call rate with a minimum of 98%. Genotypes were imputed using MACH/minimac software to the 1000 Genomes phase I version three-reference panel (all populations). From the imputed data (HRC version 1.1), we selected eight SNPs associated with ADHD at a genome-wide threshold of significance (P<10^-8) [Table SM1] (27 (link)). Four genetic variants were not included because were not available in the HRC imputations, nor were there are any variants in linkage disequilibrium, likely because the original ADHD GWAS used a custom genotyping array for psychiatric disorders. Furthermore, we constructed a genetic risk score (GRS) by multiplying the number of risk alleles by their reported odds ratio (after natural logarithm transformation) for the disease and summing this weighted allele score of each variant up into a disease risk score for ADHD.

Genotyping was performed with the Illumina 550 K, 550 K duo, and 610 quad arrays.^{13 (link)} We then removed samples with call rate below 97.5%, gender mismatch,
excess autosomal heterozygosity, duplicates or family relations and ancestry
outliers, and variants with call rate below 95.0%, failing missingness test,
Hardy–Weinberg equilibrium p-value <10⁻⁶, and
minor allele frequency <1%. Genotypes were imputed to the 1000 Genomes phase
I version 3 reference panel (all population) by using the MACH/minimac software.^{15 (link)}