hssling/genomerisklab-calibration-1kg

GenomeRiskLab PRS Calibration Distributions

RESEARCH USE ONLY — Not for clinical or diagnostic purposes.

Pre-computed PRS calibration distributions for the GenomeRiskLab platform. Each file stores percentile breakpoints (1–100) for a specific PGS Catalog score × ancestry combination, derived from 1000 Genomes Project Phase 3.

Data Format

Each calibration file is a JSON object at {pgs_id}_{ancestry}.json:

{
  "pgs_id": "PGS000018",
  "ancestry": "EUR",
  "reference_name": "1kg_EUR_v3",
  "n_samples": 2504,
  "mean": 0.412,
  "std": 0.118,
  "percentiles": {
    "1": -0.312, "2": -0.287, ..., "99": 0.891, "100": 1.204
  }
}

Ancestry Codes

Code	Population
EUR	European
AFR	African
AMR	Admixed American
EAS	East Asian
SAS	South Asian

Covered PGS IDs

Disease	PGS ID	Variants
Coronary Artery Disease	PGS000018	300
Type 2 Diabetes	PGS000036	1,000
Breast Cancer	PGS000004	313
Atrial Fibrillation	PGS000055	1,030,736
Prostate Cancer	PGS000662	451
Alzheimer's Disease	PGS000334	23
Chronic Kidney Disease	PGS001351	22
Stroke	PGS000039	90

Source Data

• 1000 Genomes Project Phase 3: IGSR
• PGS Catalog weights: pgscatalog.org

Generating Calibration Files

# From the GenomeRiskLab repo:
python scripts/build_calibration.py \
  --pgs-id PGS000018 \
  --ancestry EUR \
  --vcf-dir /data/1kg/EUR/ \
  --output-dir data/calibration/

Creator / Author

Dr Siddalingaiah H S Professor, Department of Community Medicine Shridevi Institute of Medical Sciences and Research Hospital, Tumkur, India

• Email: hssling@yahoo.com
• ORCID: 0000-0002-4771-8285
• GitHub: hssling

Disclaimer

These calibration distributions are based on publicly available reference panels and are provided for research purposes only. They do not constitute medical advice and should not be used for clinical decision-making. PRS performance varies across ancestry groups and may not generalise to all populations.