CytoImageList: S4 class for list of images

In BodenmillerGroup/cytomapper: Visualization of highly multiplexed imaging data in R

S4 class for list of images

Description

This class facilitates the handling of multiple one- or multi-channel images. It inherits from SimpleList setting elementType="Image". Therefore, each slot contains an either one- or multi-dimensional array in form of an Image object.

Usage

CytoImageList(
  ...,
  on_disk = FALSE,
  h5FilesPath = NULL,
  BPPARAM = SerialParam()
)

Arguments

...	A list of images (or coercible to a list) or individual images
on_disk	Logical indicating if images in form of HDF5Array objects (as .h5 files) should be stored on disk rather than in memory.
h5FilesPath	path to where the .h5 files for on disk representation are stored. This path needs to be defined when on_disk = TRUE. When files should only temporarily be stored on disk, please set h5FilesPath = getHDF5DumpDir()
BPPARAM	parameters for parallelised processing. This is only recommended for very large images. See MulticoreParam for information on how to use multiple cores for parallelised processing.

Details

Similar to the Image class, the first two dimensions of each entry indicate the spatial dimension of the image. These can be different for each entry. The third dimension indicates the number of channels per Image. Each entry in the CytoImageList class object must contain the same number of channels. Here, each channel represents pixel values indicating measurement intensities or in case of segmentation masks the cells' ID. The CytoImageList class therefore only supports a Grayscale colormode (see colormode) representation of each individual image.

The class further contains an elementMetadata slot that stores image-level meta information. This slot should be accessed using the mcols accessor function.

Value

A CytoImageList object

Restrictions on entry names

The CytoImageList class only supports unique entry names to avoid duplicated images. Names of a CytoImageList object can be get and set via names(x), where x is a CytoImageList object. Furthermore, only named or unnamed CytoImageList objects are allowed. Partially named objects causing empty or NA names return an error.

Coercion

Coercion to and from list, SimpleList and List:

as.list(x), as(x, "SimpleList"), as(x, "SimpleList"):

Coercion from a CytoImageList object x

as(x, "CytoImageList"):

Coercion from a list, SimpleList or List object x to anCytoImageList object

Looping

While lapply and mapply return regular list objects, endoapply and mendoapply return CytoImageList objects.

On disk representation

When setting on_disk = TRUE and specifying the h5FilesPath, images are stored on disk. To convert back to an in-memory CytoImageList object, one can call CytoImageList(on_disk_IL, on_disk = FLASE).

Author(s)

Nils Eling ([e-mail])

See Also

Image, for further image analysis tools.

SimpleList, for basics functions to handle SimpleList objects

?loadImages, for reading images into a CytoImageList object

?"CytoImageList-naming", for setting and getting image and channel names

?"CytoImageList-subsetting", for subsetting and accessor functions

Examples

# Creation of CytoImageList
u <- matrix(rbinom(100, 10, 0.5), ncol=10, nrow=10)
v <- matrix(rbinom(100, 10, 0.5), ncol=10, nrow=10)
IL1 <- CytoImageList(image1 = Image(u), image2 = Image(v))

# Coercion
as.list(IL1)
as(IL1, "SimpleList")
as(list(image1 = Image(u), image2 = Image(v)), "CytoImageList")

# On disk representation
IL1 <- CytoImageList(image1 = Image(u), image2 = Image(v),
                     on_disk = TRUE, 
                     h5FilesPath = HDF5Array::getHDF5DumpDir())

BodenmillerGroup/cytomapper documentation built on March 24, 2024, 10:21 p.m.