Skip to content

cameliaquek/maxfuse

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

55 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MaxFuse: MAtching X-modality via FUzzy Smoothed Embedding

Description

MaxFuse is a Python package for integrating single-cell datasets from different modalities with no overlapping features and/or under low signal-to-noise ratio regimes. For most single-cell cross modality integration methods, the feasibility of cross-modal integration relies on the existence of highly correlated, a priori 'linked' features. When such linked features are few or uninformative, a scenario that we call 'weak linkage', existing methods fail. We developed MaxFuse, a cross-modal data integration method that, through iterative co-embedding, data smoothing, and cell matching, leverages all information in each modality to obtain high-quality integration. A prototypical example of weak linkage is the integration of spatial proteomic data with single-cell sequencing data. For details, please refer to the paper.

This work has been led by Shuxiao Chen from Ma Lab @Upenn and Bokai Zhu from Nolan lab @Stanford.

Installation

MaxFuse is hosted on pypi and can be installed via pip. We recommend working with a fresh virtual environment. In the following example we use conda.

conda create -n maxfuse python=3.8
conda activate maxfuse
python -m pip install maxfuse

Vignettes

Example1: Protein -- RNA test run on ground-truth CITE-seq here.

Example2: Protein -- RNA test run on tissue here.

Note in cases when integrating single cell data across protein and RNA modalities, many times the nomenclature of features are different (e.g., mRNA ITGAM could be named as CD11b-1 when used as antibody). We gathered a .csv file that covers many of such naming conversions and used during the MaxFuse process. Of course, this is not a complete conversion, and users should manually add in new naming conversions if they were not included in this .csv file.

API documentation

For detailed documentation of MaxFuse API, you can visit our readthedocs page.

Code archive

The analysis presented in the manuscript was also deposited in this GitHub repository, under this folder. Note in the manuscript we used a development version of MaxFuse with slightly different grammar and can also be found there. If you require additional information on the analysis/data, please contact Zongming Ma ([email protected]).

License

MaxFuse is under the Academic Software License Agreement, please use accordingly.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 84.2%
  • Python 9.1%
  • R 6.4%
  • Shell 0.3%