BioRSP is an advanced tool developed to enhance the analysis of single-cell gene expression data within embedding spaces generated by dimensionality reduction techniques like t-SNE and UMAP. By simulating radar scanning, BioRSP systematically analyzes gene expression across various cell clusters, providing both qualitative and quantitative insights:
- It introduces Radar Scanning Plots (RSPs), which visually capture gene expression patterns from multiple angles, highlighting variations across clusters.
- BioRSP also computes metrics such as the RSP area, Root Mean Square Deviation (RMSD), and deviation scores, categorizing gene expression patterns into distinct groups.
This comprehensive approach improves the detection of cellular heterogeneity and helps distinguish between general expression patterns and key biomarker expression levels, making it a crucial tool for gene expression research.
Currently, BioRSP is available as a working package in this repository. You can download the latest stable version from the GitHub release page. More information about installing bioRSP will be provided in the future.
In the future, we plan to release BioRSP as a standalone package on PyPI for easier installation. As this project is still in its infancy, if you have any suggestions or feedback, please open an issue.
If you use BioRSP in your research, please cite the following preprint:
Detecting Data Embedding Spatial Patterns and Identifying Biomarkers with BioRSP
Zeyu Yao, Jake Y. Chen
bioRxiv 2022 Feb 1. doi: 10.1101/2022.02.01.493761.