What does this research mean for the field?

Recessive genetic models identify novel and ancestry-specific loci associated with Type 2 Diabetes that are missed by traditional additive models, explaining ancestry-related heterogeneity in allelic effects. Novelty: ClaimNovelty.SYNTHESIS. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To identify new genetic loci associated with Type 2 diabetes (T2D) using a recessive model, focusing on ancestry-specific effects.

June 7, 2026

3011-LB: Recessive GWAS Meta-analysis of 1.5M Participants Reveal Genetic Architecture of Type 2 Diabetes and Explain Ancestry-Specific Heterogeneity in Allelic Effects

Key Points

To identify new genetic loci associated with Type 2 diabetes (T2D) using a recessive model, focusing on ancestry-specific effects.
Meta-analyzed 13 T2D recessive GWAS including 300k cases and 1.2M controls (35.8% non-European).
Considered variants as recessive if they exhibited at least two-fold larger effect size under recessive compared to additive models.
Identified 129 independent associations with stronger evidence under the recessive model, including 28 novel signals.
Notable variant rs3862269 showed an odds ratio of 1.07 (P = 6 × 10-9), impacting individuals homozygous for the minor allele.
Approximately 3% of known loci demonstrated recessive effects, including HNF1A with an odds ratio of 1.15 (P = 1 × 10-15) in homozygous carriers.

Abstract

Introduction and Objective: The genetic architecture of complex traits has largely been studied using additive GWAS models that assume a linear allele effect. However, this can miss recessive signals, where increased risk is observed only among individuals homozygous for the minor allele. We aimed to identify novel loci associated with T2D under a recessive model using largest and most diverse GWAS meta-analysis. Methods: We meta-analyzed 13 T2D recessive GWAS including 300k cases and 1.2M controls (35.8% non-European ancestry). Variants were considered recessive if they showed a two-fold larger effect size or greater statistical significance under recessive model compared to additive. Results: We identified 129 independent associations with stronger statistical evidence under the recessive model, compared to additive, including 28 novel signals. Among them, a novel variant rs3862269 showed effect only in homozygous individuals (OR = 1.07, P = 6 × 10-9) and is the lead eQTL for ELAVL4 in pancreatic islets, a gene implicated in beta cell function. We also identified ancestry-specific recessive signals, such as rs12712928 near SIX3, enriched in East Asians, which showed a significant recessive effect (OR = 1.08, P = 2 × 10-13) with no effect in heterozygous (OR = 1.01, P = 0.8), which may explain the ancestry heterogeneity in allelic effects due to much lower frequency of homozygous outside East Asians. We show that approximately 3% of known loci demonstrated recessive effects. For example, at HNF1A locus, the strongest signal was a common missense variant (rs1169288) with stronger recessive association (OR = 1.11, P = 1 × 10-52; additive OR = 1.05, P = 3 × 10-41) and only effect in homozygous carriers (OR = 1.15, P = 1 × 10-15), but not in heterozygous. Conclusion: This study highlights the importance of recessive effects in T2D, identifying novel and ancestry specific associations missed by additive models, which may impact risk prediction models and explain ancestry related heterogeneity. Disclosure M. Vora: None. K. Taylor: None. J. Li: None. A. Huerta: None. S. Hsu: None. L. Saso-Jiménez: None. R. Mandla: None. C.I. Fernández Hernández: None. A. Manning: None. L. Turk: None. T. Tusié-Luna: None. J. Choi: None. S. Namba: None. M. Udler: Advisory Panel; Ended; Novo Nordisk. Research Support; Current; Novo Nordisk. M. Kurki: None. D. Carey: None. P. Palta: None. P. Natarajan: Research Support; Current; Allelica. Consultant; Current; Amgen Inc. Consultant; Ended; Broadview Ventures, Eli Lilly and Company. Consultant; Current; Foresite Labs. Consultant; Ended; Foresite Capital. Consultant; Current; Parameter Health. Other - Royalties; Current; Recora. Other - Spousal employment; Current; Vertex Pharmaceuticals Incorporated. S. Kwak: None. S. Finer: Research Support; Current; Pfizer Inc., AstraZeneca, Novo Nordisk, Takeda Pharmaceutical Company Limited, Bristol-Myers Squibb Company, GlaxoSmithKline plc., Maze Therapeutics, Merck Sharp & Dohme Corp. U. Mirshahi: None. J. Mercader: None. Funding American Diabetes Association (11-22-ICTSPM-16), National Human Genome Research Institute (U01HG011723), National Institute Of Diabetes And Digestive And Kidney Diseases - NIH (R01DK137993, R01DK140545)

Mark Helpful

Bookmark

Relay