comorbid with ADHD with sufficient literature sampling sizes. For this process, sets of genes co-mentioned with terms related to autism or schizophrenia in MEDLINE abstracts and abstract metadata were identified, using the literature mining tool FABLE (http://fable.chop.edu), the Phenopedia tool of the gene resource HuGE Navigator (http://hugenavigator.net/), and for autism queries, the autism gene database AutDB (http://www.mindspec.org/autdb.html). FABLE query terms used were ‘autism' and/or ‘schizophrenia', with or without terms relating to structural variation (see Table 1). HuGE Navigator concepts used were Autistic Disorder, Schizophrenia, and for controls, Meningioma, Astrocytoma, Parkinson Disease, and Alzheimer Disease. For AutDB, we used all genes listed as candidates as the test group. For all three resources, the sets of genes implicated in the queried diseases were tested for significantly enriched representation of ADHD CNV genes using the single-sample proportions test.
