Title
SCENT-Py: A Python toolkit for single-cell differentiation potency estimation using signaling entropy
Author
Yuhu Liang and Andrew E. Teschendorff
Affiliation
CAS Key Lab of Computational Biology, Shanghai Institute of Nutrition & Health, Chinese Academy of Sciences, 320 Yue Yang Road, Shanghai 200031, China
The signaling entropy rate (SR) quantifies the differentiation potential of a single cell by modeling its transcriptomic state as a stochastic diffusion process over a PPI network. The underlying assumption, as defined in (Teschendorff and Enver 2017). CCAT (Correlation of Connectome and Transcriptome), an ultra-fast approximation to SR (Teschendorff et al. 2021)
Here, we provide Python implementations of both the SR and CCAT algorithms. In addition, we applied several optimizations to the SR procedure to address the long runtime observed in the original R version. Based on our benchmarking, computing SR for 1 million cells on 32 cores using the Python version requires approximately 8.5 hours, representing a 4.5-fold speed improvement over the R implementation, which takes about 39 hours under the same conditions.
If you are more comfortable working in R, please refer to the link below for instructions on installing the R-based SCENT package and guidance on its usage. https://github.com/aet21/SCENT
Teschendorff, A. E. and Enver, T. (2017), 'Single-cell entropy for accurate estimation of differentiation potency from a cell's transcriptome', Nat Commun, 8, 15599.
Teschendorff, A. E., et al. (2021), 'Ultra-fast scalable estimation of single-cell differentiation potency from scRNA-Seq data', Bioinformatics, 37 (11), 1528-34.