Open, honest science, and room to collaborate
Everything behind the PenetranSci signature is built on public data and reported in full, including what didn't work. If you work on depression, inflammation or blood biomarkers, we'd like to hear from you.
The data behind the signature
Eight public whole-blood gene-expression (RNA) studies from the NCBI Gene Expression Omnibus, 779 people in total. Each passed the same quality check before use.
Simple by design
Each of six genes is compared with healthy controls in the same study; the score is the three inflammation genes minus the three lymphocyte genes. No fitted weights. The probability is calibrated across studies, and every study's result is reported separately.
Reported in full
AUC 0.61 on 7 studies the signature had never seen, above chance in 8 of 8, about the same as a standard blood count. Read the full evidence →
Open to inspect
The app's code is public on GitHub, and the app is free for academic research. The analysis code is available on request for collaborations and replication.
Research collaborations we're looking for
- Joint studies on blood immune profiles in depression and related conditions
- Sharing a cohort, especially with CRP, cytokines or treatment-response data, to test the inflammatory subgroup
- Student projects and theses in data science, neuroscience or psychiatry
- Using the app in your own research, with our support
Research updates: what we're testing next
Whether the inflammatory subgroup matches measured CRP; whether it predicts treatment response; and whether plasma proteins (proteomics) sharpen the signal.
How to cite
Delva NC. NeuroPenetrance PenetranSci immune signature (Stage 1), 2026. https://neuropenetrance.com/research.html