QUANTIFYING THE EFFECT OF COPY-NUMBER VARIANTS ON GENERAL INTELLIGENCE IN UNSELECTED POPULATIONS

Background Detection of Copy Number Variants (CNVs) are routinely performed in patients with Neurodevelopmental Disorders (NDs) and “Clinically significant” CNVs, defined as rare and large CNVs contributing to disease, are identified in 10 to 15% of patients. Effects of these clinically significant CNVs on cognitive traits have been studied for only a small number of recurrent CNVs. In addition, case-control studies are impossible to perform for > 75% of CNVs that are non-recurrent. As a result, their effects on neurodevelopment are neither characterized nor understood. Objectives: To examine the effect of CNVs on measures of Performance and Verbal Intelligence Quotient (PIQ and VIQ) in general populations. To model and predict the effect of CNVs on PIQ and VIQ using variables that characterize gene content and noncoding regions involved in CNVs. Methods We called CNVs from genotyping data with PennCNV and QuantiSNP on two cohorts drawn for the general population: Imagen (n=1804) and the Saguenay Youth Study (n=968). Rare ( Results We identified rare deletion and duplications larger than 250 kb in 10% of individuals. Both the size and gene content of rare deletions decrease IQ, eg. deletions ≥250Kb decrease IQ by 6 points (p=2.10–3). We were unable to detect a significant effect of rare duplications on IQ. For estimating the effect of all deletions on IQ, a stepwise linear model procedure converged on a model including mutation intolerance scores (pvalue 80%, p≤2.10–3). Discussion Our results suggest that the effects of deletions on general intelligence can be reliably modeled and represent a new perspective for the study of non-recurrent CNVs. These results will also help clinicians estimate the impact of non-recurrent CNVs on cognition in their patients.