Bayesian segmentation of imaging-based spatial transcriptomics data. Baysor finds cells from molecule positions and gene composition, with or without a prior segmentation. This is the native C++ release cpp-0.9.0.
Linux x86-64 — download the release binary and run on your molecule table:
curl -fLO https://github.com/kharchenkolab/Baysor/releases/download/cpp-0.9.0/baysor-0.9.0-linux-x86_64.tar.gz
tar -xzf baysor-0.9.0-linux-x86_64.tar.gz
./baysor-0.9.0-linux-x86_64/bin/baysor run -m 30 -s 8 molecules.csvDocker — from the directory containing molecules.csv:
docker run --rm -v "$PWD:/data" ghcr.io/kharchenkolab/baysor:0.9.0 run -m 30 -s 8 molecules.csvThe table needs x, y and gene columns (optional z for 3D).
30 molecules and radius 8 are examples, not universal settings.
| Setting | What to choose |
|---|---|
-m |
Minimum molecules expected in a real cell; choose for your protocol. |
-s / --scale |
Approximate cell radius in coordinate units. Alternatively, pass a prior as the second input and set --prior-segmentation-confidence (default 0.2). |
-c |
TOML config file; explicit CLI flags override it. |
-o |
Output directory (default segmentation). |
--threads |
Worker threads; physical CPU cores by default. |
- Installation — macOS / Windows binaries, requirements, Docker and source builds
- Cell segmentation — choosing parameters, using a prior and inspecting results
- Xenium workflow and examples
- Performance — run time, memory and segmentation examples
The documentation is versioned per release. For the old Julia implementation, see the migration guide and archived v0.7.1 docs.
If you use Baysor in a publication, please cite:
Petukhov V, Xu RJ, Soldatov RA, Cadinu P, Khodosevich K, Moffitt JR & Kharchenko PV.
Cell segmentation in imaging-based spatial transcriptomics.
Nat Biotechnol (2021). https://doi.org/10.1038/s41587-021-01044-w