biobatchnet 0.1.1

A deep learning framework for batch effect correction in biological data

BioBatchNet

Installation

Clone the Repository

Clone the repository to your local machine:

git clone https://github.com/Manchester-HealthAI/BioBatchNet](https://github.com/Manchester-HealthAI/BioBatchNet

Set Up the Environment

Create a virtual environment and install dependencies using environment.yml:

Using Conda:

conda env create -f environment.yml
conda activate bbn

BioBatchNet Usage

Enter BioBatchNet

cd BioBatchNet

Construct dataset

For the IMC dataset, place the dataset inside:

mv <your-imc-dataset> Data/IMC/

For scRNA-seq data, create a folder named gene_data inside the Data directory and place the dataset inside:

mkdir -p Data/gene_data/
mv <your-scrna-dataset> Data/scRNA-seq/

Batch effect correction

For IMC Data To process IMC data, run the following command to train BioBatchNet:

python imc.py -c config/IMC/IMMUcan.yaml

For scRNA-seq Data To process scRNA-seq data, modify the dataset, run the following command to train BioBatchNet:

python scrna.py -c config/IMC/macaque.yaml

CPC Usage

CPC utilizes the embedding output from BioBatchNet as input. The provided sample data consists of the batch effect corrected embedding of IMMUcan IMC data.

To use CPC, ensure you are running in the same environment as BioBatchNet.
All experiment results can be found in the following directory:

cd CPC/IMC_experiment

✅ Key Notes:

• CPC requires embeddings from BioBatchNet as input.
• Sample data includes batch-corrected IMMUcan IMC embeddings.
• Ensure the same computational environment as BioBatchNet before running CPC.

📂 Data Download Link

To use BioBatchNet for batch effect correction, you need to download the corresponding dataset and place it in the appropriate directory.

🔹 Download scRNA-seq Data

The scRNA-seq dataset is available on OneDrive. Click the link below to download:

🔗 Download scRNA-seq Data

🔹 Download IMC Data

The IMC dataset can be accessed from the Bodenmiller Group IMC datasets repository. Visit the link below to explore and download the datasets:

🔗 IMC Datasets - Bodenmiller Group

To Do List

• Data download link
• Checkpoint
• Benchmark method results

License

This project is licensed under the MIT License. See the LICENSE file for details.