scIso

An R package for quantifying and visualizing isoform-level heterogeneity in single-cell and spatial long-read RNA-seq data.
It provides a reproducible framework to compute isoform usage intercellular variability and to visualize it across conditions, and pseudotime trajectories.

---

Key features

• Quantification

  • computePIMatrix() — compute isoform Percent Inclusion (PI) from Seurat assays.
  • computePHIMatrix() — estimate isoform-level variability (φ) using moment-matched Beta parameters.

• Visualization

  • FeaturePlotPI() — project isoform PI or ratios on embeddings (UMAP, t-SNE, etc.).
  • IsoformBoxPlot() — compare isoform PI/expression across conditions with optional Wilcoxon p-values.
  • IsoformPseudotimePlot() — plot isoform dynamics along pseudotime with annotated metadata bands.

• Integration

  • Use Seurat v5 objects and compatible with most scRNA-seq long-read pipelines (Sicelore, wf-single-cell...).

---

Installation

You can install the development version from GitHub:

# install.packages("devtools")
devtools::install_github("alihamraoui/scIso")

Dependencies:
Seurat, ggplot2, Matrix, dplyr, tidyr, stringr, paletteer, patchwork, RColorBrewer, scales, rlang.
Optional: ggpubr (>= 0.6.2) for statistical comparisons in IsoformBoxPlot().

---

Development

• Written in R (≥ 4.3) using roxygen2, devtools, and testthat.
• All functions include built-in checks for reproducibility and Seurat compatibility.
• Tested with Seurat v5 .