Welcome to the new BRIE (>=2.0 or BRIE2), Bayesian Regression for Isoform Estimate, a scalable Bayesian method to accurately identify splicing phenotypes in single-cell RNA-seq experiments and quantify isoform proportions and their uncertainty.
BRIE2 supports the analysis of splicing processes at two molecular levels, either between alternative splicing isoforms or between unspliced and spliced RNAs. In either case, it returns cell-by-event or cell-by-gene matrices of PSI value and its 95% confidence interval (quantification) and the statistics for detecting DAS and DMG on each event or gene:
- Differential alternative splicing (DAS): This task is to quantify the proportions of alternative splicing isoforms and to detect DAS between groups of cells or along with a continuous covariate, e.g., pseudotime. BRIE2 is designed for two-isoform splicing events with a focus on exon skipping, but in principle also applicable for mutual exclusion, intron-retaining, alternative poly-A site, 3' splice site and 5' splice site.
- Differential momentum genes (DMG): This task is to quantify the proportions of unspliced and spliced RNAs in each gene and each cell. Similar to DAS, the DMG is a principled selection of genes that capture heterogeneity in transcriptional kinetics between cell groups, e.g., cell types, or continuous cell covariates, hence may enhance the RNA velocity analyses by focusing on dynamics informed genes.
BRIE2 is available through PyPI. To install, type the following command
line, and add -U
for upgrading:
pip install -U brie
Alternatively, you can install from this GitHub repository for the latest (often development) version with the following command line
pip install -U git+https://github.com/huangyh09/brie
In either case, add --user
if you don't have the write permission for your
Python environment.
For more instructions, see the installation manual.
- The full manual is at https://brie.readthedocs.io
- More examples and tutorials: https://github.com/huangyh09/brie-tutorials
In short, there are two steps for running BRIE2.
First, obtain cell-by-gene or cell-by-event count matrices for each isoform.
For the exon-skipping event, you can run brie-count
, which will return count
matrices and hdf5 file for AnnData.
For spliced and unspliced matrices, we listed a few options in the manual.
Then you can use brie-quant
to perform quantification of splicing ratio and
detect differential alternative splicing or differential momentum genes.
Type command line brie-count -h
and brie-quant -h
to see the full
arguments.
- Yuanhua Huang and Guido Sanguinetti. BRIE2: computational identification of splicing phenotypes from single-cell transcriptomic experiments. Genome Biology, 2021; 22(1):251.
- Yuanhua Huang and Guido Sanguinetti. BRIE: transcriptome-wide splicing quantification in single cells. Genome Biology, 2017; 18(1):123.