Cases were classified using the Autism Diagnostic Interview-Revised (ADI-R) and Autism Diagnostic Observation Schedule (ADOS) instruments and those with known karyotypic abnormalities or genetic disorders were excluded. Informed consent was obtained from all families and procedures had approval from institutional review boards. DNA was obtained from blood or buccal-swabs (73% of cases; 75% of controls) or cell-lines (22% of cases; 25% of controls) (in 5% of cases the DNA source was not identified). The 1,287 EA controls passing all QC-filters included 1,261 individuals recruited as controls for the study of addiction (SAGE)15 (link) and 26 HapMap samples (from Illumina). An additional 3,677 EA controls from three separate studies genotyped on other platforms were also used. Raw data from ASD family (Accession pending) and SAGE control (Accession: phs000092.v1.p1) genotyping are at NCBI dbGAP. CNVs were analysed using PLINK v1.0730 (link), R stats and custom scripts. Primary analyses were robust to potential systematic measurement differences between cases and controls; it was not possible to control for site but we controlled for the overall extent and number of CNVs for all burden comparisons, and obtained a consistent enriched gene count in ASD cases compared to controls.