Compute Spatial-Aware Cell Embeddings with Image Information

This function computes low-dimensional cell embeddings from a seyrat object. The method initializes cell embeddings using approximate PCA and refines them through a linear factor model nested a intrinsical conditional autoregressive model.

cellembedding_image_seurat(
  seu,
  adjm,
  assay = NULL,
  slot = "data",
  q = 10,
  approx_Phi = FALSE,
  reduction.name = "caesar",
  var.features = NULL,
  ...
)

Arguments

seu: A Seurat object containing gene expression data. The object should have variable features identified prior to running this function.
adjm: A spatial adjacency matrix representing relationships between cells or spots.
assay: A character string specifying which assay to use from the Seurat object. If NULL, the function will use the default assay set in the Seurat object.
slot: The data slot to use for feature extraction (e.g., "data", "scale.data"). Default is "data".
q: An integer specifying the number of dimensions for the reduced embeddings. Default is 10.
approx_Phi: Logical, indicating whether to use an approximate method for estimating the Phi matrix. Default is FALSE.
reduction.name: A character string specifying the name for the dimensional reduction result. Default is "caesar".
var.features: A vector of variable features (genes) to use for the analysis. If NULL, the function will automatically use the variable features stored in the Seurat object.
...: Additional arguments passed to `cellembedding_image_matrix`.

Value

The modified Seurat object with the cell embedding results stored in the specified dimensional reduction slot.

Arguments

Value

See also