https://github.com/const-ae/linear_perturbation_prediction-paper
https://github.com/const-ae/linear_perturbation_prediction-paper
Last synced: 5 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/const-ae/linear_perturbation_prediction-paper
- Owner: const-ae
- License: mit
- Created: 2024-08-28T10:05:11.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2025-07-18T11:54:44.000Z (12 months ago)
- Last Synced: 2025-09-10T05:34:08.463Z (10 months ago)
- Language: HTML
- Size: 60.8 MB
- Stars: 49
- Watchers: 3
- Forks: 9
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
README
# Code repository for *Deep learning-based predictions of gene perturbation effects do not yet outperform simple linear baselines*
This repository contains the code to reproduce our analysis from
> Deep learning-based predictions of gene perturbation effects do not yet outperform simple linear baselines.
Constantin Ahlmann-Eltze, Wolfgang Huber, Simon Anders.
_Nature Methods_ 2025; doi: https://doi.org/10.1038/s41592-025-02772-6
A copy of the code is permanently archived at https://doi.org/10.5281/zenodo.14832393.
- The **notebooks** folder contains the R scripts used for the analysis and to make the figures
- [Double Perturbation Analysis](https://htmlpreview.github.io/?https://github.com/const-ae/linear_perturbation_prediction-Paper/blob/main/notebooks/double_perturbation_analysis.html)
- [Single Perturbation Analysis](https://htmlpreview.github.io/?https://github.com/const-ae/linear_perturbation_prediction-Paper/blob/main/notebooks/single_perturbation_analysis.html)
- [Dataset Overview](https://htmlpreview.github.io/?https://github.com/const-ae/linear_perturbation_prediction-Paper/blob/main/notebooks/dataset_overview.html)
- The **benchmark** folder contains the scripts to reproduce the benchmark results
- The **benchmark/src** contains individual scripts to run each method
- The **benchmark/conda_environments** and **benchmark/renv** contain the details about the software versions
- The **benchmark/submission** contains the script to launch the scripts using my [custom](https://github.com/const-ae/MyWorkflowManager) workflow manager