The CytoMDS package implements a low dimensional visualization of a set of cytometry samples, in order to visually assess the ‘distances’ between them. This, in turn, can greatly help the user to identify quality issues like batch effects or outlier samples, and/or check the presence of potential sample clusters that might align with the experimental design.
The CytoMDS algorithm combines, on the one hand, the concept of Earth Mover’s Distance (EMD), a.k.a. Wasserstein metric and, on the other hand, the Multi Dimensional Scaling (MDS) algorithm for the low dimensional projection.
Also, the package provides some diagnostic tools for both checking the quality of the MDS projection, as well as tools to help with the interpretation of the axes of the projection.
License
The CytoMDS code is provided under GPL license version 3.0 or higher. The documentation, including the manual pages and the vignettes, are distributed under a CC BY-SA 4.0 license.
Citation
If you use CytoMDS in your research, please use the following citation:
Hauchamps, Philippe, Simon Delandre, Stephane T. Temmerman, Dan Lin, and Laurent Gatto. 2024. “Visual Quality Control with CytoMDS, a Bioconductor Package for Low Dimensional Representation of Cytometry Sample Distances.” bioRxiv. https://doi.org/10.1101/2024.07.01.601465.
or run citation("CytoMDS") to get the bibtex entry.