Exome analysis of 22,319 individuals links extremely rare copy-number variants and 22q11.21 dosage to Alzheimer risk

Journal Article (2026)
Author(s)

Olivier Quenez (Université de Rouen)

Catherine Schramm (Université de Rouen)

Kevin Cassinari (Université de Rouen)

Aude Nicolas (Sorbonne Université)

Joan Groeneveld (Vrije Universiteit Amsterdam)

Guillaume Huguet (Centre Hospitalier Universitaire Sainte-Justine Research Center)

Benjamin Grenier-Boley (LabEx DISTALZ - U1167-RID-AGE Facteurs de risque et déterminants moléculaires des maladies liées au vieillissement)

M. Hulsman (TU Delft - Electrical Engineering, Mathematics and Computer Science, Vrije Universiteit Amsterdam)

G. Bragi Walters (Amgen-deCODE Genetics)

More Authors (External organisation)

Research Group
Pattern Recognition and Bioinformatics
DOI related publication
https://doi.org/10.1016/j.ajhg.2026.07.013 Final published version
More Info
expand_more
Publication Year
2026
Language
English
Research Group
Pattern Recognition and Bioinformatics
Journal title
American Journal of Human Genetics
Issue number
10
Volume number
113
Pages (from-to)
2099-2117
Page Views
22
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

Copy-number variants (CNVs) are major contributors to human disease. In Alzheimer disease (AD), APP duplications cause autosomal-dominant forms, but the role of CNVs in non-monogenic AD remains poorly characterized. We analyzed rare CNVs (frequency <1%) from 22,319 exomes (4,150 early-onset AD [EOAD, ≤65 years], 8,519 late-onset AD [LOAD], 9,650 unaffected control subjects) using harmonized calling and quality control. After identifying 17 individuals with a pathogenic CNV, we performed exome-wide and gene-set burden analyses. EOAD-affected individuals showed increased burdens of rare CNVs affecting coding genes, particularly deletions in AD-related genes. Integrated loss-of-function (LoF) analysis gathering short truncating variants with deletions showed that ABCA1 (odds ratio [OR] = 5.77 [95% confidence interval 2.25; 17.06], p = 0.0002) and ABCA7 deletions contribute to this deletion burden (OR = 2.29 [1.44; 3.65], p = 0.0006), while CTSB LoF alleles appear as candidates (OR = 5.03 [1.50; 20.71], p = 0.0089). We then performed exome-wide gene-level dosage analysis and highlighted 18 genes across five loci with a false discovery rate of <10%, including the 22q11.21 central region, where deletions were restricted to EOAD (including one de novo event) and duplications were enriched in control individuals, with intermediate frequencies in LOAD. We narrowed this locus to the SCARF2 - KLHL22 - MED15 region after integrating short truncating variants. Replication in 33,977 affected individuals and 362,322 control subjects confirmed association for 22q11.21 dosage with exome-wide significance (OR
SCARF2 = 0.34 [0.21; 0.53]; mega- p value = 5.52 × 10
−7). SCARF2 overexpression significantly increased amyloid-β uptake, congruent with duplication-associated decreased AD risk. We conclude that rare coding CNVs in a proportion of AD-associated genes and 22q11.21 deletions, including some found in DiGeorge syndrome, increase AD risk. Conversely, we identify 22q11.21 duplication as a strong AD-risk-decreasing factor.