Rare variant genetic architecture of human cortical organization

856 

Abstract Submission 

Kuldeep Kumar¹, Sayeh Kazem¹, Zhijie Liao¹, Jakub Kopal², Guillaume Huguet³, Thomas Renne¹, Martineau Jean-Louis³, Zhe Xie¹, Zohra Saci³, Laura Almasy⁴, David Glahn⁵, Tomas Paus⁶, Guillaume Dumas¹, Carrie Bearden⁷, Paul Thompson⁸, Richard Bethlehem⁹, Varun Warrier¹⁰, Sébastien Jacquemont³

¹CHUSJ Research Center, University of Montreal, Montreal, Quebec, ²McGill University, Montreal, Quebec, ³CHU Saint-Justine, Montreal, QC, ⁴Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, ⁵Department of Psychiatry, Harvard Medical School, Boston, MA, ⁶Université de Montréal, Montreal, Quebec, ⁷University of California at Los Angeles, Los Angeles, CA, ⁸Imaging Genetics Center, Keck School of Medicine of University of Southern California, Los Angeles, CA, ⁹Autism Research Centre, Department of Psychiatry, University of Cambridge, Cambridge, United Kingdom, ¹⁰University of Cambridge, Cambridge, Cambridgeshire

Kuldeep Kumar  
CHUSJ Research Center, University of Montreal
Montreal, Quebec

Sayeh Kazem  
CHUSJ Research Center, University of Montreal
Montreal, Quebec

Zhijie Liao  
CHUSJ Research Center, University of Montreal
Montreal, Quebec

Jakub Kopal  
McGill University
Montreal, Quebec

Guillaume Huguet  
CHU Saint-Justine
Montreal, QC

Thomas Renne  
CHUSJ Research Center, University of Montreal
Montreal, Quebec

Martineau Jean-Louis  
CHU Saint-Justine
Montreal, QC

Zhe Xie  
CHUSJ Research Center, University of Montreal
Montreal, Quebec

Zohra Saci  
CHU Saint-Justine
Montreal, QC

Laura Almasy  
Perelman School of Medicine, University of Pennsylvania
Philadelphia, PA

David Glahn  
Department of Psychiatry, Harvard Medical School
Boston, MA

Tomas Paus  
Université de Montréal
Montreal, Quebec

Guillaume Dumas  
CHUSJ Research Center, University of Montreal
Montreal, Quebec

Carrie Bearden, PhD  
University of California at Los Angeles
Los Angeles, CA

Paul Thompson, PhD  
Imaging Genetics Center, Keck School of Medicine of University of Southern California
Los Angeles, CA

Richard Bethlehem  
Autism Research Centre, Department of Psychiatry, University of Cambridge
Cambridge, United Kingdom

Varun Warrier  
University of Cambridge
Cambridge, Cambridgeshire

Sébastien Jacquemont  
CHU Saint-Justine
Montreal, QC

The development and organization of the human cortex are highly heritable [Hibar (2015), Elliott (2018), Grasby (2020)]. Genome-wide association studies (GWAS) have identified common variants influencing the global and regional cortical phenotypes derived using magnetic resonance imaging (MRI) [Grasby (2020), Smith (2021), Warrier (2023)]. Studies focusing on a very small set of rare variants, previously associated with neuropsychiatric disorders, have shown large effects on cortical structures [Modenato (2021)]. However, genome-wide studies have not been conducted and the rare variant architecture of the cerebral cortex remains unknown. Here we aimed to map the effects of rare variants, genome-wide, on twelve MRI-derived phenotypes, measured globally and regionally, in general population cohorts. We also compared the phenotype burden genetic correlations with common variant genetic correlations to gain insights into the relationship between common and rare variant architectures.

We analyzed structural and diffusion MRI-derived phenotypes and all copy-number-variants (CNV) >50 kilobases from 40,000 UK Biobank participants [Smith (2021), Moreau (2023)] and 8,000 ABCD participants [Hagler (2019)]. Our analysis focused on seven macrostructural measures derived from T1w MRI: cortical thickness (CT), surface area (SA), volume (Vol), folding index, intrinsic curvature index, mean curvature (MC), and Gaussian curvature; and five microstructural measures derived from diffusion MRI: fractional anisotropy, mean diffusivity (MD), isotropic volume fraction (ISOVF), intracellular volume fraction, and orientation diffusion index [Warrier (2023)].
Since the current sample size remains underpowered to detect gene-level or variant-level associations for most rare variants, we used a functional burden association test. The latter aggregates the variants that disrupt genes involved in a given biological function in each individual. Genes were assigned to biological functions using previously published cell-type marker genes [Wagstyl (2022)], Gene Ontology terms, and spatial transcriptomics [Hawrylycz (2012)].
The burden association analysis is a linear model that estimates the mean effect size on MRI phenotypes of genes within each gene-set (and corresponding biological function) of interest. Analyses are performed, for deletions and duplications separately, and are adjusted for age, sex, site, and ancestry. We estimated burden genetic correlations between pairs of MRI phenotypes and between variant-types by correlating the effect sizes across all biological functions by adapting a previously published approach [Weiner (2023)].

Out of 24,768 burden associations (12 phenotypes x 1032 gene sets x 2 variant type), 437 showed FDR significance. Vol, SA, and CT showed the strongest association with gene dosage, while ISOVF, MD, and MC showed much weaker associations. The CNV burden genetic correlations were positive and negative across 58% and 42% pairs of traits, respectively. These CNV burden genetic correlations were concordant with those previously published using common-variants (r=0.84, p<1e-4). The effect sizes of deletions and duplications were negatively correlated for the majority of phenotypes, suggesting that deletions and duplications show preferential effects across MRI traits and biological functions. Analyses at the regional level showed significant associations across many functional gene sets. The CNV burden genetic correlations across brain regions were concordant with those previously reported for common variants.

Our study revealed that the common and rare variant architectures of human cortical MRI phenotypes are concordant, with most MRI phenotypes preferentially sensitive to either deletions or duplications. Because deletions and duplications have large negative and positive effects on gene expression (respectively), our analysis provides insight into the effects of changes in transcription on MRI phenotypes.

Genetic Association Studies ¹

Genetic Modeling and Analysis Methods ²

Cortex

Cortical Layers

STRUCTURAL MRI

Structures

Other - Copy number variants; Rare variant architecture; Burden Association