Sashimi-style coverage plot for a splice event

Draws junction arcs scaled by read count over a schematic gene structure. Arcs are coloured by role: inclusion (blue) vs exclusion (red). Works in both junction mode and transcript mode. Optionally faceted by a cell metadata column.

Draws junction arcs scaled by aggregate read count over a schematic gene structure. Arcs are coloured by role: inclusion (blue) vs exclusion (red).

Usage

PlotSashimi(
  object,
  event_id,
  cells = NULL,
  group_by = NULL,
  arc_scale = c("sqrt", "linear", "log"),
  colours = c(inclusion = "#4393c3", exclusion = "#d6604d"),
  title = NULL,
  ...
)

# S4 method for class 'MatisseObject'
PlotSashimi(
  object,
  event_id,
  cells = NULL,
  group_by = NULL,
  arc_scale = c("sqrt", "linear", "log"),
  colours = c(inclusion = "#4393c3", exclusion = "#d6604d"),
  title = NULL,
  ...
)

Arguments

object: A MatisseObject with a PSI assay computed.
event_id: Character. A single event ID. Run rownames(GetSeurat(obj)[["psi"]]) to list available IDs; typical SE format is "SE:chr1:1201-2999:3201-4999:+".
cells: Character vector of cell barcodes to aggregate over. Default: all cells.
group_by: Character. Column in Seurat meta.data to facet by. Default: NULL (all cells pooled).
arc_scale: Character. How to scale arc height to read count: "sqrt" (default), "linear", or "log".
colours: Named character vector with elements "inclusion" and "exclusion" giving arc colours.
title: Character. Plot title. Defaults to event_id.
...: Additional arguments (see PlotSashimi).

Value

A ggplot object.

Details

In junction mode each arc corresponds to an individual junction with its own read count. In transcript mode the SE event_id is parsed to derive junction coordinates; inclusion and exclusion counts come from the "counts" and "exclusion" layers of the PSI assay.

Supported event types in transcript mode: SE (skipped exon) and RI (retained intron). Junction mode supports all event types since coordinates are auto-parsed from junction IDs.

Note on transcript-mode SE arcs: transcript-level counting aggregates reads to events, not to individual junctions. The total inclusion-event count is therefore split evenly across the two SE inclusion arcs in the plot. The two arcs may have differed in reality; use junction-mode input if you need per-junction read counts.