DANCE documentation

DANCE is a Python toolkit to support deep learning models for analyzing single-cell gene expression at scale. It includes three modules at present:

1. Single-modality analysis

2. Single-cell multimodal omics

3. Spatially resolved transcriptomics

Our goal is to build up a deep learning community for single-cell analysis and provide GNN based architectures for users to further develop in single-cell analysis.

Getting started

To install the DANCE package, first make sure you have correctly set up dependencies such as PyTorch, PyG, and DGL. Then, simply install the DANCE package from PyPI via

pip install pydance

Alternatively, use our installation script provided in our Github repository to setup a dance conda environment with all necessary dependencies as well as DANCE.

git clone git@github.com:OmicsML/dance.git && cd dance
source install.sh cu117  # other options are [cpu,cu118]

Finally, you can checkout some example scripts we provided to reproduce some of the experiments from the original papers.

Modules

• Single modality tasks
  • Cell type annotation
  • Clustering
  • Imputation
• Multi modality tasks
  • Joint embedding
  • Modality matching
  • Modality prediction
• Spatial transcriptomic tasks
  • Spatial domain
  • Cell type deconvolution

API Reference

• dance.data
• dance.datasets
• dance.transforms
• dance.transforms.graph

Indices and tables

• Index

• Module Index

• Search Page