Visualize isoform percent inclusion (PI), ratios and distributions on embeddings

This function visualizes the relative usage (PI, Percent Inclusion) of two isoforms across cells on a low-dimensional embedding stored in a Seurat object. It computes per-cell PI for each isoform, their ratio and log2-ratio, adds these values to the object metadata, and displays:

Usage

FeaturePlotPI(
  obj,
  iso1,
  iso2,
  reduction,
  assay = "ISO",
  slot = "counts",
  name = NULL,
  add_ratio = TRUE,
  min_total = 0,
  col = "PI",
  bins = 30
)

Arguments

obj: A Seurat object.
iso1: Isoform ID for the first isoform (must be a row name in the assay).
iso2: Isoform ID for the second isoform (must be a row name in the assay).
reduction: Name of the dimensional reduction to use (e.g. "umap").
assay: Assay name containing isoform counts. Default: "ISO".
slot: Slot to use in the assay (e.g. "counts", "data"). Default: "counts".
name: Optional base name used to construct metadata column names.
add_ratio: Logical; if TRUE, add ratio and log2-ratio to metadata. Default: TRUE.
min_total: Numeric threshold: cells with iso1 + iso2 < min_total are set to NA for PI. Default: 0.
col: Suffix used for PI-based metadata/label naming. Default: "PI".
bins: Number of bins for the PI histogram. Default: 30.

Value

A patchwork object combining three feature plots and one histogram.

Details

PI of isoform 1 on the embedding,
PI of isoform 2 on the embedding,
log2(ratio iso1 / iso2) on the embedding,
histogram of the PI distribution for isoform 1.

Examples

if (FALSE) { # \dontrun{
library(Seurat)
library(Matrix)

set.seed(42)

mat <- Matrix(
  rpois(200, lambda = 10),
  nrow = 2,
  sparse = TRUE,
  dimnames = list(
    c("GeneA..Tx1", "GeneA..Tx2"),
    paste0("Cell", 1:100)
  )
)

obj <- CreateSeuratObject(counts = mat, assay = "ISO")

emb <- matrix(
  rnorm(200),
  ncol = 2,
  dimnames = list(colnames(obj), c("UMAP_1", "UMAP_2"))
)

obj[["umap"]] <- CreateDimReducObject(
  embeddings = emb,
  key = "UMAP_",
  assay = "ISO"
)

FeaturePlotPI(obj, "GeneA..Tx1", "GeneA..Tx2", reduction = "umap")
} # }