This score predicts the likelihood of having blonde hair color. Blonde hair results from lower levels of the dark pigment eumelanin in hair follicles. It is most common in populations of Northern European descent and is influenced by several genes, particularly variants near the MC1R and KITLG genes.
Score Details
This score comes from the PGS Catalog, an open repository of published polygenic scoring models used in genomic research.
- PGS ID
- PGS001093
- Score Name
- GBE_BIN_FC1001747
- Reported Trait
- Hair color, blonde
- Development Method
- snpnet
- Number of Variants
- 6,970
- Original Genome Build
- GRCh37
- Weight Type
- NR
- Release Date
- October 21, 2021
Source Publication
Significant sparse polygenic risk scores across 813 traits in UK Biobank.
Development Samples
Polygenic risk scores are developed against large reference datasets. This score was built using the UK Biobank, a biomedical cohort with genetic and health data from roughly 500,000 participants — large enough to identify reliable genetic associations.
| Ancestry | Sample Size | Country | Cohort |
|---|---|---|---|
| European(white British ancestry) | 269,201 | UK | UKB |
Ancestry Distribution
Development (269,201 individuals)
Evaluation (3 sample sets)
Performance Evaluations
Polygenic risk scores are usually developed and validated in populations of European ancestry. Predictive accuracy can vary across ancestries because of differences in genetic architecture, linkage disequilibrium patterns, and allele frequencies. For binary traits like disease risk, AUROC measures how well the score separates cases from controls (0.5 = random, 1.0 = perfect). Values substantially below the best-performing ancestry are highlighted to flag reduced transferability.
| AncestryThe broad ancestral population of the evaluation sample. | NNumber of individuals in the evaluation sample. | R²Proportion of variance explained by the full model (PGS + covariates). Higher is better. | PGS R²Proportion of variance explained by the PGS alone, without covariates. Isolates the predictive power of the genetic score. | AUROCArea Under the Receiver Operating Characteristic curve. Measures how well the score discriminates cases from controls. 0.5 = random, 1.0 = perfect. |
|---|---|---|---|---|
| European | 24,865 | 0.3570 | 0.3592 | 0.8629[0.8561, 0.8697] |
| South Asian | 7,652 | 0.0703 | 0.0184 | 0.7223[0.5638, 0.8808] |
| European | 67,286 | 0.3220 | 0.3240 | 0.8464[0.8422, 0.8506] |
Covariates: age, sex, UKB array type, Genotype PCs
Downloads & Links
PGS obtained from the Catalog should be cited appropriately, and used in accordance with any licensing restrictions set by the authors. See EBI Terms of Use (https://www.ebi.ac.uk/about/terms-of-use/) for additional details.