Skip to content

theislab/sc_mae

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Masked Autoencoder in Single-Cell Genomics

This repository provides a simple setting to train a Masked Autoencoder (MAE) on single-cell genomics data with a random masking strategy. The provided code is designed to work with a smaller scale adata object that fits into memory.

Contents

  • data.py: Module for loading and preprocessing single-cell genomics data.
  • Masking.ipynb: Jupyter notebook demonstrating the random masking strategy.
  • models.py: Contains the implementation of the Masked Autoencoder.
  • train.py: Script for training the Masked Autoencoder model.
  • train.sh: Bash script for executing the training process.

System Requirements

  • Python 3.10
  • Dependencies listed in requirements.txt

Usage

  1. Clone the repository:
git clone https://github.com/theislab/sc_mae.git
  1. Install the required dependencies:
pip install -r requirements.txt
  1. Prepare the data:
  • Download the sample data from the publication mentioned in the citation section or use your own processed adata object.
  1. Execute the training script:
bash train.sh

Demo

To apply this code, follow these steps:

  1. Download Sample Data: You can download the adata object from the publication mentioned in the citation section or use your own processed h5ad object.
  2. Prepare Data: If you are using your own data, make sure it is preprocessed and compatible with the provided code. Otherwise, follow the data loading and preprocessing steps in data.py.
  3. Train the Model: Execute the training script train.py by running bash train.sh. Adjust the hyperparameters and configurations as needed in the script.

Citation

This repository is a part of a larger project and serves as a simplified demo. If you use this code in your research, please cite the following paper:

Delineating the Effective Use of Self-Supervised Learning in Single-Cell Genomics

Link to the paper

Link to the repository

If you use the sample data in your research, please cite the following paper:

COMBATdb: a database for the COVID-19 Multi-Omics Blood ATlas

Link to the paper

Acknowledgments

  • The sample data used in this project is sourced from the COMBATdb.

Contribution

Contributions to improve this codebase are welcome. Please fork the repository and submit a pull request with your changes.

License

This project is licensed under the MIT License - see MIT License.

Please refer to the main repository for more detailed information and a more elaborate analysis.

Authors

sc_mae was written by Till Richter.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published