Abstract: TH-PO0106
Conditional Logistic Ancestry-Inclusive Regression Analysis (CLAIRA) Enables Ancestry-Inclusive Rare Disease Genome-Wide Association Study Using UK Biobank Controls
Session Information
- Integrative Genomics and Multi-Omics in Noncystic Genetic Kidney Disease
October 22, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Genetic Diseases of the Kidneys
- 1202 Genetic Diseases of the Kidneys: Non-Cystic (Complex and Non-Cystic Monogenic)
Authors
- Doctor, Gabriel, University College London, London, England, United Kingdom
- Sadeghi-Alavijeh, Omid, University College London, London, England, United Kingdom
- Voinescu, Catalin D., University College London, London, England, United Kingdom
- Tzoumkas, Konstantinos, University College London, London, England, United Kingdom
- Carter, Matthew A., University College London, London, England, United Kingdom
- Demeneva, Darja, University College London, London, England, United Kingdom
- Smith, Joanna K., University College London, London, England, United Kingdom
- Carr, Edward, University College London, London, England, United Kingdom
- Downie, Mallory Lorraine, McGill University, Montreal, Quebec, Canada
- Levine, Adam P., University College London, London, England, United Kingdom
- Chan, Melanie M.Y., Imperial College London, London, England, United Kingdom
- Stanescu, Horia, University College London, London, England, United Kingdom
- Gale, Daniel P., University College London, London, England, United Kingdom
Group or Team Name
- Centre for Genetics and Genomics
Background
Rare diseases contribute disproportionately to end-stage kidney disease. Rare disease genome-wide association studies (GWAS) are limited by small case numbers and diverse genetic ancestries. GWAS methods commonly rely on exclusion of genetically admixed individuals and restriction to the cohort's major ancestral grouping to address population stratification, but this may reduce power and equity. We developed CLAIRA, an ancestry-inclusive association testing framework utilising genetic similarity matching and conditional logistic regression (CLR). We applied it to Idiopathic Nephrotic Syndrome (INS).
Methods
Fourteen INS cohorts were assembled from international collaborators. Cases and some controls were genotyped on Illumina MEGA and GSA arrays. Genetic principal components (PCs) were calculated jointly with UK Biobank (UKB) to facilitate using UKB as a source of external controls. CLAIRA matches each case to unique controls within a PC-distance threshold, prioritising case inclusion. Virtual genotyping arrays were generated from phased UKB whole-genome sequence data to maximise marker intersection with each genotyping array and a novel method applied to address batch effects. Samples were imputed against a UKB 500k-sample reference panel. CLAIRA performs CLR on PLINK2 dosage or hard-call data, treating each case and its matched controls as a stratum. It assumes an additive genetic model and supports covariate adjustment and genotype residualisation on stratum-centred PCs. Execution time is comparable to PLINK2 association testing.
Results
CLAIRA matched 1719 INS cases to 2-10 controls, mostly from UKB. By comparison only 1530 cases could be included in an ancestry-specific meta-analysis. In two highly admixed array cohorts, genomic inflation was substantially reduced compared with standard logistic regression (MEGA λGC 1.2 vs 2.6; GSA <1.1 vs 2.6; both without PC-adjustment). Meta-analysis of CLAIRA with PC-residualised genotypes (λGC 1.08) and the largest published INS study reinforced nine loci not previously reaching genome-wide significance, including one not detected using ancestry-specific analyses alone.
Conclusion
CLAIRA enables ancestry-inclusive GWAS in rare kidney disease while maintaining acceptable genomic control. This approach increases usable case numbers and may improve equity and locus discovery in under-represented populations.
Funding
- Private Foundation Support