This project contains the Python code used in the paper "AI-driven high-throughput droplet screening of cell-free gene expression." The repository includes multiple sets of code, each serving a distinct purpose in the study:
- Data pre-processing
- Neural Network Model
- XGBoost
- Transfer Learning Combined with XGBoost Code
- Ubuntu 20.04 LTS
- Windows 10/11
- macOS Big Sur and later
- Python: Version 3.8 or higher
- MATLAB: Version R2023a or higher
/Clustering Cleaning & TSNE/cleaning.mlx:Matlab Live scripts are used for data preprocessing, including data clustering, cleaning, and T-SNE./Combination Optimization/method2_1.h5: A pretrained model used to test demo, compiled with tensorflow keras./Combination Optimization/combination_optimization.ipynb: Jupyter notebook for model building and training./XGBoost & Transfer Learning/main.py: The main script that executes the core algorithm./XGBoost & Transfer Learning/code_demo: Contains demo and related code required for the demo.
You can install all dependencies with the following command:
pip install -r requirements.txt
- Python: 3.8, 3.9
- TensorFlow: 2.5.0, 2.6.0
- scikit-learn: 0.24.2
- Matlab: R2023a
Estimated Time: The installation typically takes 5-10 minutes on a standard desktop computer with a reliable internet connection.
Place your dataset in the appropriate directory.
Run the Clustering Cleaning & TSNE/cleaning.mlx file to perform data preprocessing.
Run the Combination Optimization/combination_optimization.ipynb file to build and train the model.
Run the XGBoost & Transfer Learning/main.py file to build and train the model.
You can run the following demo code:
-
/Clustering Cleaning & TSNE/cleaning.mlx -
/XGBoost & Transfer Learning/code_demo/demo_transfer_learning.ipynb -
/XGBoost & Transfer Learning/code_demo/demo_XGBoost_model.ipynb
After running the codes, you will obtain relevant results and evaluations of the model.
This project is open-sourced under the MIT License. For more details, see the LICENSE file.