msBayesImpute (Python version)

msBayesImpute is a versatile framework for handling missing values in mass spectrometry (MS) proteomics data.
It integrates probabilistic dropout models with Bayesian matrix factorization in a fully data-driven manner,
allowing it to account for both missing at random (MAR) and missing not at random (MNAR) patterns.

This repository contains the Python implementation of msBayesImpute, built on Pyro, a probabilistic programming language.
The R version is available here: msBayesImpute (R package).

---

Repository structure

msbayesimputepy/
├── data/                       # Example dataset (HeLa cell line proteomics data)
├── msbayesimputepy/            # Python implementation of msBayesImpute
├── msbayesimputepy.egg-info/   # Metadata for the Python package
├── dist/                       # Pre-built Python wheel package
├── vignettes/                  # Example usage (see quick_guide_python.ipynb)
├── requirements.txt            # Package dependencies
└── README.md

---

Installation

Install the Python package from the pre-built wheel in the dist/ folder:

pip install dist/msbayesimputepy-0.2.0-py3-none-any.whl

---

Getting started

• See the Jupyter notebook in vignettes/quick_guide_python.ipynb for a quick start.
• Example dataset: provided in the data/ folder (HeLa cell line proteomics).

---

Citation

If you use msBayesImpute in your research, please cite:
He J, et al. bioRxiv (2025). msBayesImpute: A Versatile Framework for Addressing Missing Values in Biomedical Mass Spectrometry Proteomics Data

---