signatomics
Python tools for Olink proteomics data: read, quality control, bridging and sample metadata, with
every step reported.
signatomics is a family of Python tools for finding biomarker signatures in omics data. The
first module, signatomics.olink, works with Olink (PEA) proteomics data. More modules will
build on it later.
📖 Documentation: https://signatomics.elenastamatelou.com
Status: in development (version 0.0.1). The Olink module reads data, controls its quality,
bridges plates and adds your sample metadata; signatomics.stats compares groups and correlates.
Functions may still change before the first release.
What it does today
import signatomics.olink as so
data = so.read_npx(so.example_file()) # a synthetic example NPX file; your own: so.read_npx("my_npx_file.csv")
so.qc_report(data) # look at quality: your data stays as it is
clean = so.prepare(so.example_file()) # read, quality control, bridging and sample metadata in one call
Each step also exists on its own, with every choice visible: read_npx,
qc_report → apply_qc, bridge_check → bridge_plates, add_sample_info.
- Reading: NPX files in long format and older wide (NPX Manager) files; .csv, .txt, .xlsx and .parquet;
European decimal commas, "No Data" cells, TRUE/FALSE written in other languages.
- Quality control: a quality check and charts that leave your data as it is, then your choices
applied, with every removal listed.
- Bridging: per-protein plate correction, with the detection limits shifted along, and an
independent PCA check.
- Sample metadata: adds your sample metadata (demographics, clinical data, ...) and reports every ID
without a match.
- Statistics: the analysis table (
to_wide: one row per sample), then compare_groups and
correlate on every assay, with effect sizes, 95% intervals and an FDR; plot_volcano.
- Charts: interactive Plotly charts in one style, colour-blind safe; sample IDs hidden by default.
Install
pip install "signatomics[olink] @ git+https://github.com/ElenaStamatelou/signatomics"
Requires Python 3.10 or newer. See Start in 5 minutes.
The family
| Module | Data | Status |
|---|
signatomics.olink | Olink (PEA) proteomics: Target 96/48, Explore, Explore HT, Reveal | in development |
signatomics.stats | compare groups, correlate: every assay, effect sizes, FDR | in development |
| core | one standard table, combining data types, biomarker signature with nested cross-validation | planned |
| more modules | e.g. ELISA, microbiome | ideas |
Repository layout
| Folder | What |
|---|
src/signatomics/ | the package |
tests/ | automatic tests (pytest) |
analyses/ | worked analyses of public data (Jupyter notebooks) |
website/ | the documentation website (MkDocs Material) |
docs/ | function reference pointer, notes on the example datasets |
Public datasets stay outside this repository; the only data inside are the small, synthetic example
files of the package. See docs/datasets.md.
Author
Elena Stamatelou
signatomics is an independent project. It is not affiliated with or endorsed by Olink or Thermo
Fisher Scientific. Olink is a trademark of Thermo Fisher Scientific.
