Quantifying the Effect of Copy-number Variants on General Intelligence in Unselected Populations
1. Université de Montréal; 2. CHU Sainte-Justine; 3. The Hospital for Sick Children, University of Toronto, Toronto, Ontario Canada, M5G 1X8; 4. Institute of Psychiatry, King’s College London, United Kingdom; 5. Lady Davis Institute for Medical Research, Jewish General Hospital; 6. McGill University; 7. School of Psychology, University of Nottingham, Nottingham, UK; 8 Rotman Research Institute, University of Toronto, Toronto, ON, Canada; 9. Institut Pasteur, Paris, France; 10. CNRS URA 2182 “Genes, synapses and cognition”, Paris, France; 11. Université Paris Diderot, Sorbonne Paris Cité, Paris, France
Introduction: With the routine implementation in the clinic of whole genome chromosomal microarrays (CMAs), “clinically significant” Copy Number Variants (CNVs) (defined as rare variants contributing significantly to disease) are currently identified in 10 to 15 % of patients referred for a Neurodevelopmental Disorders (NDs). However, the effects of CNVs on essential cognitive traits such as general intelligence have been studied for only a small number of CNVs. Examination of cohorts of individuals who carry the same CNV (such as 16p11.2, 22q11.2) to study the impact on cognitive traits is informative, but over 75% of “clinically significant” CNVs are non-recurrent and observed only once or a few times in patients. Their effect on cognition and behavioral traits are neither characterized nor quantified.
Objectives: To establish the first estimates of the effect of intermediate size, large & rare (1/1000) non-recurrent CNVs on cognition using general population cohorts. Also, to investigate variables that contribute the most to IQ variance and model the effects of CNVs on general intelligence.
Preliminary results: We called CNVs on two general population cohorts of European ancestry: Imagen (n=1804) and Saguenay Youth Study (SYS, n=977). We identified rare CNVs (<1/1000 in Data Genome Variants) larger than 100Kb based on Illumina genotyping chip data in Imagen (263 deletions and 282 duplication) and SYS (160 deletions and 244 duplication) and show that deletions ≥250Kb significantly decrease IQ by approximately 4 points (p=2.10-3). On the other hand, large duplications do not show any effect on this cognitive trait.
We then sought to identify characteristics explaining the effect of CNVs on IQ. Gene and regulatory content was annotated for all CNVs based on scores of intolerances to mutation, temporal and tissue expression, metrics of protein-protein interactions. Then we fit multivariate models where the most predictive independent variables were selected from among the above characteristics by AIC. The best model was then validated on an independent set of data containing CNVs with robust empirical data.
Our preliminary results suggest that the effects of non-recurrent CNVs on cognitive traits can be reliably modeled. This will help clinicians to estimate the impact of non-recurrent CNVs on cognition in their patients referred for a ND or a psychiatric condition.