ProteoForge Manuscript Analysis

Status	License	Language	Release	Zenodo	Citation

This repository contains the code, analyses, and rendered figures supporting the ProteoForge manuscript. It includes real-data benchmarks, simulation studies, and an application to a hypoxia study. This repo is the analysis snapshot.

The scripts used here (ProteoForge/) are not packaged; they are simply a collection of functions developed alongside the manuscript. Some features — especially plotting and printing helpers — were added ad hoc and are specific to this analysis. A standalone Python package implementing the ProteoForge methodology is currently under development and will be made publicly available upon stable release.

Repository Layout

Top-level folders and their purpose:

• ProteoForge/ — core Python scripts used in analyses (parsers, processing, modelling, clustering, classifiers).
• Benchmark/ — scripts and notebooks for benchmark analyses (R and Python).
• NSCLC/ — notebooks, data and figures for the hypoxia/NSCLC application.
• Simulation/ — simulation scripts, notebooks and utilities used to evaluate methods.
• src/ — auxiliary Python library used by some scripts (utilities, plotting helpers, tests).
• requirements.txt, setup_project.sh, setup_project.ps1, setup_env.R — environment and setup helpers.

The setup utilities ensure you have venv and renv folders created with the required dependencies. They setup the environment for both R and Python analyses to facilitate reproducibility across OSes.

Notes on data and outputs:

• Raw and derived data/figures are not committed. Place raw inputs under the appropriate */data/input/ folders; scripts/notebooks will write to */data/ and */figures/ (see folder READMEs).
• A snapshot of the repository with input data, and the html renders of all notebooks, is available at Zenodo: 10.5281/zenodo.17795845.

Environment Setup (Cross-Platform)

Use the provided setup scripts to configure both Python (venv) and R (renv + pak). R 4.5.0 or newer is required for the R environment.

Linux / macOS (bash):

git clone https://github.com/LangeLab/ProteoForge_Analysis.git
cd ProteoForge_Analysis
bash setup_project.sh

Windows (PowerShell):

git clone https://github.com/LangeLab/ProteoForge_Analysis.git
cd ProteoForge_Analysis
./setup_project.ps1

If R is not on PATH, install it from CRAN and rerun the setup command, or run Rscript setup_env.R after installation. To activate the Python environment later, use source .venv/bin/activate (Linux/macOS) or ./.venv/Scripts/Activate.ps1 (PowerShell).

Run Steps

Entry points for reproducing analyses and figures:

• Notebooks: Benchmark/*.ipynb, Simulation/*.ipynb, NSCLC/*.ipynb.
• Scripts (Python): Benchmark/04-runProteoForge.py, Simulation/04-runProteoForge.py.
• Scripts (R): Benchmark/01-DataProcessing.R, Benchmark/02-runCOPF.R, Benchmark/03-runPeCorA.R, plus analogous scripts in Simulation/.

Each notebook/script documents its required inputs and outputs. Place raw inputs under the corresponding */data/input/ directory before running. Outputs will be written under */data/ and */figures/.

Reproducibility Notes

• R environment: managed with renv; run via setup_project.sh/setup_project.ps1 or Rscript setup_env.R. Required R version: >= 4.5.0.
• Python environment: requirements.txt lists dependencies; the setup scripts create .venv and install the requirements.
• Data locations: inputs are expected under */data/input/; outputs are written to */data/ and */figures/. Large files are not tracked in git.
• Software vs analysis: this repository is the analysis snapshot. A standalone Python package is under development and will be made publicly available upon stable release.

Citations

Please cite the manuscript and the analysis snapshot when using this work.

• Manuscript (preprint):
  • ProteoForge: An Imputation-Aware Framework for Differential Proteoform Discovery in Bottom-Up Proteomics. bioRxiv. Posted December 16, 2025. (under review)
• Analysis snapshot (this repository): use the Zenodo record and select the version matching the git tag you used.
  • "Snapshot of Benchmarking and Showcasing ProteoForge for Proteoform Deconvolution from Peptide Level Data. Version 1. Zenodo. 10.5281/zenodo.17795845."

License

This repository is licensed under CC BY-NC 4.0: see license.