Interactive spatial feature viewer

FeatureViewer opens up an interactive shiny application where one can zoom and pan the H&E image while overlaying and color selected features. It is also possible to add new categorical features or modify existing categorical features with a lasso tool.

Usage

FeatureViewer(
  object,
  slot = "data",
  datadir = NULL,
  selected_features = NULL,
  sampleIDs = NULL,
  host = "127.0.0.1",
  port = 8080L,
  custom_color_palettes = NULL,
  categorical_colors = NULL,
  container_width = 800,
  container_height = 800,
  nCores = detectCores() - 1,
  launch.browser = TRUE,
  verbose = TRUE
)

Arguments

object: A Seurat object created with semla
slot: A slot to use for Assay data
datadir: A directory containing spatial data and image tiles (see ExportDataForViewer)
selected_features: A character vector of features to select for the viewer. These features need to be present in the 'default assay' or they can be names of dimensionality reduction vectors.
sampleIDs: An integer vector of section IDs to use include in the viewer. All sections will be used by default. Note that if you have exported the data with ExportDataForViewer, you need to make sure that the sampleIDs for FeatureViewers matches the sampleIDs used for ExportDataForViewer.
host: A string with a valid IPv4 or IPv6 address to listen on, Defaults to localhost "127.0.0.1"
port: An integer that indicates the port to listen on.
custom_color_palettes: A named list of color vectors to use as custom color palettes
categorical_colors: A named list of character vectors with color names. The name of each character vector should correspond to a categorical variable in the meta.data slot of the Seurat object. Each character vector should be named where each name corresponds to a label of the category.
container_width, container_height: Set height and width of container
nCores: Number of cores to use for threading passed to TileImage. Only used if datadir is unavailable.
launch.browser: Select browser to use. See runApp for details
verbose: Print messages

Value

A Seurat object created with semla

Details

The viewer requires a tiled H&E image along with some additional image data to work. By default, the function will attempt to create these files and export them to a temporary directory, but it is also possible to export the files before running the app and provide the path to the data directory with datadir (see ExportDataForViewer).

A more detailed tutorial can be found on the semla website. You can also get more detailed instructions by pressing the help icon in the app.

Selecting features

FeatureViewer automatically selects the features in the 'default assay' and dimensionality reduction vectors. You can override this selection with selected_features, but note that these have to be present in the 'default assay' or in the dimensionality reduction slot. This is to avoid naming conflicts when multiple assay are present which contains identical identifiers, for instance 'Spatial' and 'SCT'.

Non-finite values

If there are non-finite values present, e.g. NA or NaN, a warning will be thrown and these values will be replaced with a 0. Alternatively, you can remove spots from the Seurat object which contain non-finite values before launching the viewer.

Examples

A tutorial can be found on our [package website](https://ludvigla.github.io/semla/). Go to tutorials -> Feature Viewer

Author

Ludvig Larsson