rob-norm 🦭🧬

Python implementation of the RobNorm R package for robust normalization of quantitative omics data.

Features

• Identical normalization results to the original R implementation.
• Fast and efficient implementation using NumPy (57 ms vs 113 ms in R for 5000x200 data on an M1 MacBook Air)

Installation

Install via pip:

pip install rob-norm

Usage

import pandas as pd
from rob_norm import rob_norm

# load your data into a DataFrame (rows = features, columns = samples)
data = pd.read_csv('./data/simulated_measurements.txt', index_col=0, sep='\t')

# alternatively, simulate data using the provided function
# sim_dat_fn(row_frac, col_frac, mu_up, mu_down, n, m, nu_fix=True, seed=None):

# perform a normalization operation provided by the package
results = rob_norm(data, gamma_0=0.5, tol=1e-4, step=200)

normalized_data = results['norm_data']

# verify against reference results
df_results = pd.read_csv('./data/simulated_measurements_normalized.txt', index_col=0, sep='\t')

assert np.allclose(df_results, normalized_data) # == True

Development

Clone the repository and use uv for development and testing:

git clone https://github.com/Tom-Julux/rob_norm
cd rob_norm

uv run pytest
uv build
uv publish

License

The project is licensed under the GNU Lesser General Public License v3.0.

See the LICENSE file for full licensing information.

The file data/robnorm.r contains R code used to generate reference results for testing and was copied and adapted from the original R package RobNorm.