RNA-Sieve is a method for the deconvolution of bulk cell samples via single-cell RNA expression data.
Our work has been published in Genome Research. The manuscript is also available on bioRxiv.
If you find RNA-Sieve useful, please cite our work at (Google Scholar):
Erdmann-Pham, D. D., Fischer, J., Hong, J., & Song, Y. S. (2021).
A likelihood-based deconvolution of bulk gene expression data using single-cell references.
Genome Research, gr-272344.
For Python 3, we recommend that you install rnasieve via pip.
$ pip3 install rnasieve
For example usage, please reference the example Jupyter notebook for Python 3 usage, or the Mathematica notebook for Mathematica usage.
The core algorithm is called find_mixtures/findMixtures which takes in a vector of bulk expressions to be deconvolved and reference matrices of means, variances, and sample counts.
Currently, only the Python library takes allows for multiple bulks to be jointly deconvolved with a single set of reference matrices.