noise_decomp

A super small Python package that implements the dual-reporter decomposition of intrinsic and extrinsic noise from paired single-cell measurements.

Install (editable)

pip install -e .

---

Usage

from noise_decomp import noise_decomp

res = noise_decomp(r, g)
print(res)

---

CLI

After installing (editable) the package you get a small CLI noise-decomp:

Compute from comma-separated values:

noise-decomp --r 1,2,3 --g 1.1,2.1,3.1

Compute from files (CSV or whitespace separated):

noise-decomp --rfile path/to/r.csv --gfile path/to/g.csv

Use --no-normalize to disable mean normalization and --ddof to set the delta degrees of freedom for variance/covariance.

---

Cython (fast) extension

To build the Cython extension for best performance:

pip install cython numpy
python build_cython.py

The Python API will automatically use the fast Cython backend if available.

---

Examples

Run the tiny demo script:

python examples/demo.py

---

References

• Elowitz MB, Levine AJ, Siggia ED, Swain PS. Stochastic Gene Expression in a Single Cell. Science (2002).

• Raser JM, O’Shea EK. Control of Stochasticity in Eukaryotic Gene Expression. Science (2004).