Human biobanks connect deeply characterized participant data with biospecimens, electronic health records, imaging, and molecular measurements. Biobank Atlas is a selective guide to the public GitHub repositories that make those resources easier to access, analyze, reproduce, and govern.
This is a catalog of code and documentation—not participant-level data. Access to controlled biobank data still requires approval from the relevant study.
- What belongs here
- Official biobank and cohort repositories
- Cross-biobank networks and standards
- Landmark public analysis resources
- Biobank-scale analysis software
- Repository at a glance
- Contributing
- Responsible use
A repository is included when it is public, clearly relevant to human biobanks, and meets at least one of these impact signals:
- 🏛️ Official — maintained by a biobank, national cohort, or research infrastructure.
- 📄 Published — supports a peer-reviewed methods or resource paper.
- 🌍 Reference resource — released a widely used biobank analysis or public results collection.
- 🧰 Reusable — provides a maintained method or workflow designed for biobank-scale data.
GitHub stars are deliberately not an inclusion threshold: official infrastructure and specialist scientific software can be highly consequential while serving a small developer community. One-off paper scripts, generic bioinformatics packages, private repositories, mirrors, and repositories that contain only a dataset citation are out of scope.
Official repositories from UK Biobank, centered on its cloud-based Research Analysis Platform (RAP).
| Repository | What it provides | Primary language | Signal |
|---|---|---|---|
| UKB-RAP-Notebooks-Access | Practical notebooks for finding, accessing, and extracting UK Biobank data on RAP. | HTML / notebooks | 🏛️ Official |
| UKB-RAP-Notebooks-Genomics | Examples for array, exome, and whole-genome data on RAP. | Jupyter Notebook | 🏛️ Official |
| UKB-RAP-Imaging-ML | Imaging and machine-learning examples, including scalable image processing. | Jupyter Notebook | 🏛️ Official |
| UKB-RAP-Workflows | Multi-stage WDL workflows designed to run on RAP. | WDL | 🏛️ Official |
| SNP-filtering | Notebook workflow for filtering genotyping data by variant or genomic region. | Shell / notebooks | 🏛️ Official |
| UKB-Git-Audit-Tool | Scans Git history for accidental exposure of potentially sensitive UK Biobank content. | Python | 🏛️ Official |
Open infrastructure from the US National Institutes of Health's All of Us Research Program.
| Repository | What it provides | Primary language | Signal |
|---|---|---|---|
| workbench | Source for the secure Researcher Workbench used to analyze the program's Curated Data Repository. | Java | 🏛️ Official |
| raw-data-repository | APIs and services for the program's participant-facing raw data repository. | Python | 🏛️ Official |
| curation | Data curation and quality workflows used to build research releases. | Python | 🏛️ Official |
| data-browser | Source for the public Data Browser. | TypeScript | 🏛️ Official |
| workbench-snippets | Reusable R and Python notebook snippets for analyses inside Workbench. | Jupyter Notebook | 🏛️ Official |
| ukb-cross-analysis-demo-project | Demonstration of a harmonized analysis spanning All of Us and UK Biobank. | Jupyter Notebook | 🏛️ Official · cross-biobank |
FinnGen combines Finnish biobank samples with national longitudinal health-register data and releases summary results publicly.
| Repository | What it provides | Primary language | Signal |
|---|---|---|---|
| finngen-documentation | Versioned documentation for public FinnGen releases, methods, endpoints, and downloads. | Markdown | 🏛️ Official |
| finemapping-pipeline | Production WDL pipeline for statistical fine-mapping of FinnGen GWAS results. | WDL | 🏛️ Official · reusable |
| saige-pipelines | WDL workflows used to run SAIGE association analyses at FinnGen scale. | WDL | 🏛️ Official · reusable |
| regenie-pipelines | WDL workflows for REGENIE association analyses. | WDL | 🏛️ Official · reusable |
| phenotype-matching | Maps FinnGen endpoints to UK Biobank/PheCode phenotypes using ICD definitions. | Python | 🏛️ Official · cross-biobank |
| META_ANALYSIS | Tools for multi-study GWAS meta-analysis and harmonization. | Python | 🏛️ Official · reusable |
| risteys | Code behind FinnGen's disease-endpoint and disease-trajectory browser. | Elixir | 🏛️ Official · reference resource |
Public tools from Our Future Health, the UK's large prospective health research program.
| Repository | What it provides | Primary language | Signal |
|---|---|---|---|
| tre-example-notebooks | Example analyses for the program's trusted research environment (TRE). | Jupyter Notebook | 🏛️ Official |
| ofh-synthetic-data-generator | Generates synthetic tabular data with an Our Future Health-like schema for development and testing. | Python | 🏛️ Official |
| tre-package-access | Machine-readable list of packages approved for restricted TRE projects. | Markdown | 🏛️ Official · governance |
| Repository | Network or standard | What it provides | Primary language | Signal |
|---|---|---|---|---|
| globalbiobankmeta/flagship | Global Biobank Meta-analysis Initiative (GBMI) | Reproducible analysis code for GBMI's multi-ancestry, multi-biobank flagship study. | R | 🌍 Reference resource |
| globalbiobankmeta/PRS | GBMI | Code for cross-biobank polygenic risk score analyses. | R | 📄 Published |
| BBMRI-ERIC/miabis | BBMRI-ERIC | MIABIS terminology and information models for describing and exchanging biobank metadata. | TypeScript / schemas | 🏛️ Official · standard |
| BBMRI-ERIC/negotiator | BBMRI-ERIC | Open-source system for structured access negotiation around samples and human data. | Java | 🏛️ Official · infrastructure |
| BBMRI-ERIC/directory-scripts | BBMRI-ERIC | Operational and export tools for the European biobank Directory. | Python | 🏛️ Official · infrastructure |
These repositories are not maintained by a biobank operator, but they produced highly reused public analyses of major biobank datasets.
| Repository | Biobank | What it provides | Primary language | Signal |
|---|---|---|---|---|
| Nealelab/UK_Biobank_GWAS | UK Biobank | QC, phenotype processing, association code, and documentation for the Neale Lab UKB GWAS release. | Python | 🌍 Reference resource |
| atgu/ukbb_pan_ancestry | UK Biobank | Code for the multi-ancestry Pan-UK Biobank GWAS resource and its analyses. | Python / R | 📄 Nature Genetics (2024) |
| Nealelab/ukb_exomes | UK Biobank | Main pipeline for single-variant and gene-based analysis across thousands of UKB exome phenotypes. | Python | 📄 Cell Genomics (2022) |
| Nealelab/ukb_common | UK Biobank | Shared phenotype and analysis code supporting the Genebass exome resource. | Python | 🌍 Reference resource |
| Nealelab/UKBB_ldsc | UK Biobank | Reproducible LD-score regression analysis and results-site generation for UKB GWAS. | R | 🌍 Reference resource |
| Repository | Best for | Primary language | Evidence of impact |
|---|---|---|---|
| rgcgithub/regenie | Fast whole-genome regression, single-variant tests, and rare-variant aggregate tests in large cohorts. | C++ | 📄 Nature Genetics (2021) |
| saigegit/SAIGE | Case-control imbalance, related samples, and scalable single-variant and gene-based tests. | C++ / R | 📄 Nature Genetics (2018) |
| hail-is/hail | Distributed manipulation and analysis of very large genomic datasets in cloud or cluster environments. | Python / Scala | 🧰 Used by UKB, gnomAD, Genebass, and Pan-UKB |
| chrchang/plink-ng | Efficient genotype QC, transformation, scoring, and association with PLINK 2. | C / C++ | 🧰 Foundational large-scale genetics toolkit |
| projectglow/glow | Genomic processing and GWAS on Apache Spark and Databricks. | Scala / Python | 📄 Bioinformatics (2021) |
| GeneticAnalysisinBiobanks/GRAB | Robust GWAS methods for binary, survival, longitudinal, and other complex biobank traits. | C++ | 🧰 Purpose-built for biobank-scale association |
| Repository | Best for | Primary language | Evidence of impact |
|---|---|---|---|
| genepi/nf-gwas | Portable, highly parallel GWAS workflows with REGENIE and interactive reporting. | Nextflow | 📄 NAR Genomics and Bioinformatics (2024) |
| ikmb/gwas-qc | Single-command QC and association testing across multi-cohort or biobank-scale studies (BIGwas). | Python / Nextflow | 📄 GigaScience (2021) |
| MRCIEU/PHESANT | Automated phenome-wide scans and consistent transformation of UKB phenotypes. | R | 📄 International Journal of Epidemiology (2018) |
| lcpilling/ukbrapR | Building reproducible R workflows on the current UK Biobank RAP. | R | 🧰 RAP-focused, actively maintained |
| kenhanscombe/ukbtools | Exploring and reshaping legacy, locally downloaded UKB filesets. | R |
| Repository | Best for | Primary language | Evidence of impact |
|---|---|---|---|
| statgen/pheweb | Building interactive browsers for hundreds or thousands of GWAS; deployed by FinnGen and other biobanks. | Python | 📄 Bioinformatics (2019) |
| FINNGEN/ld_server | Serving linkage-disequilibrium calculations against FinnGen reference panels. | Python | 🏛️ FinnGen production infrastructure |
| Area | Repositories | Start here |
|---|---|---|
| Official biobank infrastructure | 22 | UKB RAP access notebooks, All of Us Workbench, FinnGen documentation |
| Cross-biobank networks and standards | 5 | GBMI flagship, MIABIS |
| Landmark public analyses | 5 | Neale Lab UKB GWAS, Pan-UKB, UKB exomes |
| Reusable analysis software | 13 | REGENIE, SAIGE, Hail, nf-gwas |
The catalog was last reviewed on 23 August 2026. Entries are checked for a public GitHub repository, a clear connection to human biobanks, useful documentation, and at least one impact signal defined above. Repository activity and ownership can change, so every entry should be reassessed periodically.
This list aims for high-value coverage, not an impossible claim of absolute completeness. Important biobanks without meaningful public GitHub repositories are not represented. A repository's inclusion is not an endorsement of its scientific conclusions, security model, or fitness for clinical use.
Contributions are welcome. Before opening a pull request, confirm that the repository:
- Is public and contains substantive code, workflows, schemas, or technical documentation.
- Directly supports a named human biobank, cross-biobank initiative, or analysis at biobank scale.
- Has a clear license or reuse statement whenever possible.
- Has an official, peer-reviewed, reference-resource, or strong reuse signal.
- Does not expose participant-level or otherwise controlled data.
Use this template in a pull request:
- **Repository:** https://github.com/owner/repository
- **Biobank or network:**
- **One-sentence description:**
- **Primary language:**
- **Impact evidence:** official ownership, DOI, public deployment, or adoption
- **Maintenance status:** active / stable / legacyKeep descriptions factual and compact. Add a DOI for published software, label archived or legacy tools, and place the entry in the narrowest applicable section. Please do not rank projects by stars or submit repositories solely because they analyzed a biobank once.
- Follow each biobank's data-access agreement, publication policy, and participant-withdrawal requirements.
- Never commit participant-level data, credentials, access tokens, field exports, or small-cell results to a public repository.
- Treat example pipelines as starting points: verify genome build, ancestry handling, phenotype definitions, relatedness, consent scope, and software versions for your study.
- Biobank-derived models and associations may not transfer across populations. Evaluate representation, calibration, and potential harms before downstream or clinical use.
The curated text in this repository is available under CC0 1.0. Every linked project retains its own license and terms of use.