Map the similarity class of every pixel

This function visualizes the spatial distribution of gene expression similarity by classifying pixels into four categories: below threshold, similar, higher in the first dataset and higher in the second dataset. The plot is drawn with the pixel geometries of rasterized SpatialExperiment objects (colData() column geometry, which needs the sf package), otherwise with points at the spatial coordinates.

Usage

pixelClass(input, gene, assayName = NULL)

Arguments

input

A list. Results from spatialSimilarity(). This includes the similarity table, log-transformed pixel data, and analysis parameters.

gene

Character. The name of the gene to visualize.

assayName

A character string or numeric specifying the assay in the Spatial Experiment to use. Default is NULL: the assay that spatialSimilarity() used. The classification comes from input, so the plot does not depend on it.

Value

A ggplot2 spatial plot displaying classified pixels, where:

blue

Pixels classified as similar (within the fold-change threshold).

yellow

Pixels with greater expression in dataset X than Y.

red

Pixels with greater expression in dataset Y than X.

grey

Pixels with gene expression below the threshold in both experiments.

The plot title includes the gene name and its similarity score.

See also

plotGene(), the standardized figure of a gene in a compareSpatial() result, which includes this map.

Examples

data(speKidney)
##### Rasterize to get pixels at matched spatial locations #####
rastKidney <- SEraster::rasterizeGeneExpression(speKidney,
                assay_name = 'counts', resolution = 0.2, fun = "mean",
                square = FALSE)
s <- spatialSimilarity(list(A = rastKidney$A, B = rastKidney$B))
pixelClass(s, "Gene")