Sample spatial gene visualization grid

This function visualizes the gene expression stored in assays(spe) or any continuous variable stored in colData(spe) for a set of samples at the spot-level using (by default) the histology information on the background. To visualize clusters (or any discrete variable) use vis_grid_clus().

vis_grid_gene(
  spe,
  geneid = rowData(spe)$gene_search[1],
  pdf_file,
  assayname = "logcounts",
  minCount = 0,
  return_plots = FALSE,
  spatial = TRUE,
  viridis = TRUE,
  height = 24,
  width = 36,
  image_id = "lowres",
  alpha = NA,
  cont_colors = if (viridis) viridisLite::viridis(21) else c("aquamarine4",
    "springgreen", "goldenrod", "red"),
  sample_order = unique(spe$sample_id),
  point_size = 2,
  auto_crop = TRUE,
  na_color = "#CCCCCC40",
  is_stitched = FALSE,
  ...
)

Arguments

spe: A SpatialExperiment-class object. See fetch_data() for how to download some example objects or read10xVisiumWrapper() to read in spaceranger --count output files and build your own spe object.
geneid: A character() specifying the gene ID(s) stored in rowData(spe)$gene_search or a continuous variable(s) stored in colData(spe) to visualize. For each ID, if rowData(spe)$gene_search is missing, then rownames(spe) is used to search for the gene ID. When a vector of length > 1 is supplied, the continuous variables are combined according to multi_gene_method, producing a single value for each spot.
pdf_file: A character(1) specifying the path for the resulting PDF.
assayname: The name of the assays(spe) to use for extracting the gene expression data. Defaults to logcounts.
minCount: A numeric(1) specifying the minimum gene expression (or value in the continuous variable) to visualize. Values at or below this threshold will be set to NA. Defaults to 0.
return_plots: A logical(1) indicating whether to print the plots to a PDF or to return the list of plots that you can then print using plot_grid.
spatial: A logical(1) indicating whether to include the histology layer from geom_spatial(). If you plan to use ggplotly() then it's best to set this to FALSE.
viridis: A logical(1) whether to use the color-blind friendly palette from viridis or the color palette used in the paper that was chosen for contrast when visualizing the data on top of the histology image. One issue is being able to differentiate low values from NA ones due to the purple-ish histology information that is dependent on cell density.
height: A numeric(1) passed to pdf.
width: A numeric(1) passed to pdf.
image_id: A character(1) with the name of the image ID you want to use in the background.
alpha: A numeric(1) in the [0, 1] range that specifies the transparency level of the data on the spots.
cont_colors: A character() vector of colors that supersedes the viridis argument.
sample_order: A character() with the names of the samples to use and their order.
point_size: A numeric(1) specifying the size of the points. Defaults to 1.25. Some colors look better if you use 2 for instance.
auto_crop: A logical(1) indicating whether to automatically crop the image / plotting area, which is useful if the Visium capture area is not centered on the image and if the image is not a square.
na_color: A character(1) specifying a color for the NA values. If you set alpha = NA then it's best to set na_color to a color that has alpha blending already, which will make non-NA values pop up more and the NA values will show with a lighter color. This behavior is lost when alpha is set to a non-NA value.
is_stitched: A logical(1) vector: If TRUE, expects a SpatialExperiment-class built with visiumStitched::build_spe(). http://research.libd.org/visiumStitched/reference/build_spe.html; in particular, expects a logical colData column exclude_overlapping specifying which spots to exclude from the plot. Sets auto_crop = FALSE.
...: Passed to paste0() for making the title of the plot following the sampleid.

Value

A list of ggplot2 objects.

Details

This function prepares the data and then loops through vis_gene() for computing the list of ggplot2 objects.

Examples


if (enough_ram()) {
    ## Obtain the necessary data
    if (!exists("spe")) spe <- fetch_data("spe")

    ## Subset to two samples of interest and obtain the plot list
    p_list <-
        vis_grid_gene(
            spe[, spe$sample_id %in% c("151673", "151674")],
            spatial = FALSE,
            return_plots = TRUE
        )

    ## Visualize the spatial adjacent replicates for position = 0 micro meters
    ## for subject 3
    cowplot::plot_grid(plotlist = p_list, ncol = 2)
}
#> 2024-12-16 21:53:36.241884 loading file /github/home/.cache/R/BiocFileCache/10535bcd9c3d_Human_DLPFC_Visium_processedData_sce_scran_spatialLIBD.Rdata%3Fdl%3D1

Arguments

Value

Details

See also

Examples