Bering is a deep learning algorithm for simultaneous molecular annotation and cell segmentation in single-cell spatial transcriptomics data. It builds on top of torch_geometric and scanpy, from which it inherits modularity and scalability. It provides versatile models that leverages the spatial coordinates of the data, as well as pre-trained models across spatial technologies and tissues.

Visit our documentation for installation, tutorials, examples and more.

• Identify background and real signals in noisy spatial transcriptomics data.
• Identify cell annotations for transcripts on single-cell spatial data.
• Efficiently cell segmentation with cell annotations.
• Build and fine-tune pre-trained model on new data using transfer learning.

Install Bering via PyPI by running:

pip install Bering

or via Conda as:

conda install -c conda-forge Bering

Please refer to our manuscript Jin, Zhang et al. (2023) for more details.

We are happy about any feedback! If you have any questions, please feel free to contact [email protected], [email protected]. Find more research in Shu_Jian_Lab.