GWAS of Hospital Frailty Risk Score (HFRS) in FinnGen and UK Biobank

GWAS summary statistics for "Large-scale genome-wide analyses with proteomics integration reveal novel loci and biological insights into frailty" (Mak et al. 2025, Nature Aging).
Citation: Mak, J.K.L., Qin, C., Krüger, M. et al. Large-scale genome-wide analyses with proteomics integration reveal novel loci and biological insights into frailty. Nat Aging 5, 1589–1600 (2025). https://doi.org/10.1038/s43587-025-00925-y
Methods: Full details are described in the original paper. Briefly, we calculated the Hospital Frailty Risk Score (HFRS) based on 109 weighted ICD-10 codes according to a previously described protocol (Gilbert et al., Lancet, 2018) in FinnGen release 12 (discovery; n=500,737) and UK Biobank (replication; n=407,463). In FinnGen, we performed GWAS using SAIGE. We included variants (N = 21,294,561) with minor allele frequency > 0.01%, Hardy–Weinberg P value > 1 × 10−9 and imputation INFO score ≥ 0.9. The models were adjusted for birth year, sex and the first ten PCs. We performed variant mapping and annotation using the Variant Effect Predictor in the standard FinnGen GWAS annotation pipeline. In UK Biobank, we performed GWAS using fastGWA for participants of white British ancestry. We first constructed a genetic relationship matrix (GRM) using a filtered set of common SNPs and then fit a linear-mixed model for ~9.8 million variants adjusted for birth year, sex, genotyping array and the first ten PCs. The GWAS was conducted for (1) the original HFRS, and (2) HFRS excluding dementia (sensitivity analysis), given that dementia has the highest weight in the HFRS definition.
List of GWAS datasets:
"FinnGen_hfrs_gwas.txt.gz": GWAS summary statistics (SAIGE) for the HFRS in FinnGen (n=500,737). Models were adjusted for birth year, sex and first 10 ten PCs."FinnGen_hfrswodementia_gwas.txt.gz": GWAS summary statistics (SAIGE) for the HFRS without dementia in FinnGen (sensitivity analysis; n=500,737). Models were adjusted for birth year, sex and first 10 ten PCs."UKB_hfrs_gwas.txt.gz": GWAS summary statistics (fastGWA) for the HFRS in UK Biobank (n=407,463). Models were adjusted for birth year, sex genotyping array, and first 10 ten PCs."UKB_hfrswodementia_gwas.txt.gz": GWAS summary statistics (fastGWA) for the HFRS without dementia in UK Biobank (sensitivity analysis; n=407,463). Models were adjusted for birth year, sex genotyping array, and first 10 ten PCs.