Exome sequencing of the Autism NIMH Controls (AUT, 322 samples), for the Swedish Schizophrenia control data (SCZ, 503 samples) and IHCS data (IHCS, 84 samples) was carried out at the Broad Institute14–18. We only used samples ascertained as controls in the AUT and SCZ data (i.e. no presence of disease). Exons were captured using the Agilent 38Mb SureSelect v2 Libraries and sequenced using either an Illumina HiSeq2000 or Illumina GenomeAnalyzerII instrument. All samples met the criterion of >90% of targeted bases having >10x coverage and >80% of targeted bases having >20x coverage. Reads were mapped to hg19 using BWA and processed with Picard and GATK (see URLs). The SCZ samples were genotyped on the Affy 5.0 or 6.0 platforms. Genotype data across all samples (SCZ, AUT and IHCS, 909 in total) was merged with SNPs filtered by missing data and departure from Hardy-Weinberg equilibrium. Genotype likelihoods obtained using GATK28 software were passed to Beagle in windows of 1Mb with 250Kb to impute all SNPs identified as polymorphic in the haplotypes of the European 1000 Genomes Project phase 1 data. 103,977
