Spit For Science: Using Kids From The Community To Uncover The Genomics of Childhood Psychiatric Disorders

Abstract

Abstract Background Most common disorders, including Attention-Deficit Hyperactivity Disorder (ADHD) and Obsessive-Compulsive Disorder (OCD), are likely extremes of traits that are widely distributed in the general population. People are more or less restless, inattentive, impulsive, obsessional and compulsive; only the extreme of these traits demarcate a clinical diagnosis. Thus, quantitative psychiatric traits, which often share polygenic risk with their relevant diagnosis, will help us discover genes for disorders. Quantitative cognitive traits that might mediate part of the genetic risk of psychiatric disorders (i.e., cognitive endophenotypes) are often less heterogeneous than a diagnosis which could help reduce noise to improve power for gene discovery. Based on these alternative approaches, we designed the Spit for Science study to examine the genomics of childhood psychiatric disorders using behavioural traits and a cognitive endophenotype relevant to ADHD and OCD in a large population-based sample of youth. Methods Spit for Science is a sample of 17,263 children (6-17 years of age) seen at a local science museum with measures of psychiatric traits and DNA from saliva. Quantitative data on ADHD traits using the Strengths and Weaknesses of ADHD and Normal Behavior (SWAN) scale, obsessive-compulsive (OC) traits using the Toronto Obsessive-Compulsive scale (TOCS) and response inhibition using the Stop-Signal task was collected. We genotyped unrelated Caucasians (n=5,636) using Illumina HumanCoreExome beadchips. For common variant analyses, we conducted standard quality control and used hypothesis-driven genome-wide association analyses (GWAS-HD) to test biological hypotheses for each trait. For the GWAS-HD, permutation tests were used to test the whole gene set for association with each trait respectively. Copy number variants (CNVs) were called using three algorithms and only CNV calls detected by two algorithms and at least 10kb in size were retained. We examined CNV burden for each trait individually. Results We have identified the first genome-wide finding for OC traits (rs7856850; p = 2.9 x 10-8, β = 0.13) replicated in an independent sample of youth screened for OCD. Genes involved in central nervous system development were significantly associated with both OC traits and response inhibition. Polygenic risk score analyses are forthcoming. ADHD traits were associated with increased CNV deletions while response inhibition was associated with more CNV duplications and longer CNV deletions. Discussion Several promising findings have emerged from the Spit for Science study. We identified the first genome-wide finding for OC traits, identified a biological pathway involved in both OC traits and response inhibition, and demonstrated the relationship between rare CNVs with ADHD traits and response inhibition. Results to date demonstrate the feasibility, power and utility of using quantitative-trait based and cognitive endophenotype approaches in community samples to help uncover the underlying genetic architecture of psychiatric disorders.