The MatisseObject S4 class
Source:R/MatisseObject-class.R, R/MatisseObject-methods.R
MatisseObject-class.RdThe central data structure for Matisse. It wraps a Seurat
object and adds isoform-resolved splicing layers. All per-cell data – junction
counts, PSI values, transcript counts, and QC metrics – live inside the
embedded Seurat object as named assays (Assay5) or cell metadata
(meta.data). Nothing is duplicated outside the Seurat object.
Details
Two input modes are supported, set automatically at construction:
"junction"Short-read mode. Raw junction counts (from STARsolo) are stored as
Assay5("isoform")(junctions x cells). CallCalculatePSIto compute PSI values."transcript"Long-read mode. Transcript isoform counts (from Bagpiper, FLAMES, or LIQA) are stored as
Assay5("isoform")(transcripts x cells). CallCalculatePSIto compute PSI values.
After CalculatePSI, PSI values are stored as
Assay5("psi") (splice events x cells).
Functions
show(MatisseObject): Display a summary of aMatisseObject.x[[i: Access cell metadata or Seurat slots via[[. Returns metadata columns as bare vectors (matching Matisse convention), then falls back to the embedded Seurat object's[[for assays, reductions, etc. The explicit metadata check is load-bearing — Seurat's own[[returns metadata columns wrapped in a 1-column data.frame.
Slots
seuratA
Seuratobject. Contains all per-cell data: gene expression, splice assays ("isoform","psi"), cell metadata (QC metrics, cluster labels), and dimensionality reductions.input.modeCharacter.
"junction"for short-read (STARsolo junction counts) objects;"transcript"for long-read (Bagpiper / FLAMES / LIQA transcript counts) objects.versionCharacter string recording the Matisse version used to create the object.
miscNamed list for user-defined extra data.