Understand your cells,
layer by layer
Matisse computes percent spliced in (PSI) for each splicing event in each cell and stores it as an assay in your Seurat object, next to gene expression: the same cells, the same clusters, the same UMAP.
What you can discover
Questions Matisse is built to answer
Cell-type-specific splicing
Do two cell types include this exon at different rates, and by how much?
Long reads or short reads
Start from per-cell transcript counts from a long-read quantifier such as Bagpiper, or from STARsolo splice-junction counts from 10x short reads.
Works with your existing setup
Matisse layers on top of Seurat: your clusters, UMAP, and cell labels stay intact
Short-read RNA (10x) STAR / STARsolo junction count matrix
Long-read / isoform Bagpiper FLAMES LIQA
Event annotations SUPPA2 generateEvents
Installation
install.packages("remotes")
remotes::install_github("avisrilab/Matisse")