This organisation is where the analysis gets built: pipelines, project templates, tutorials for the methods we use most, and forks of the community tools we build on. Most repositories are private while the work is in progress.

Persister Cell Biology Group hosts the code accompanying our publications, one repository per paper.

No data in git. Raw and processed data stay on institutional storage and are released through GEO/EGA at publication. Repositories hold code and lightweight summary tables only.

Environments are declared. renv.lock for R, environment.yml for Python — every figure traces back to a script, a commit and an environment.

Reports read tables, not objects. Heavy steps run on the HPC and export small files, so any analysis can be re-rendered anywhere, by anyone, at any time.

One report per project. A self-contained HTML document with a fixed structure: workflow, cohort, quality control, objectives, conclusions.