imcdatasets

Documentation is available at: https://bodenmillergroup.github.io/imcdatasets

Introduction

The imcdatasets package is an extensible resource containing a set of publicly available and curated Imaging Mass Cytometry datasets. Each dataset consists of three data objects:

1. Single cell data in the form of a SingleCellExperiment class object.
2. Multichannel images formatted into a CytoImageList class object.
3. Cell segmentation masks formatted into a CytoImageList class object.

These formats facilitate accession and integration into R/Bioconductor workflows. The data objects are hosted on Bioconductor's ExperimentHub platform.

Requirements

The imcdatasets package requires R version >= 4.2. It builds on data objects contained in the SingleCellExperiment and cytomapper packages. These packages must, therefore, be installed (see below).

Installation

The release version of imcdatasets can be installed by following standard Bioconductor package installation procedures:

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("imcdatasets")

The development version can be installed from GitHub using devtools:

if (!requireNamespace("devtools", quietly = TRUE))
    install.packages("devtools")
devtools::install_github("BodenmillerGroup/imcdatasets", build_vignettes = TRUE)

Installing the dependencies (if not already done):

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install(c("SingleCellExperiment", "cytomapper"))

Loading imcdatasets in your R session:

library(imcdatasets)

Usage

Detailed information on how to access the datasets is available in the imcdatasets vignette, which is available here: https://bodenmillergroup.github.io/imcdatasets/articles/imcdatasets.html.

The vignette can also be viewed directly in R:

vignette("imcdatasets")

Details

The imcdatasets package provides quick and easy access to published and curated imaging mass cytometry datasets. Each dataset consists of three data objects that can be retrieved individually:

1. Single cell data in the form of a SingleCellExperiment class object: This object contains cell-level expression values and metadata. The rowData contain marker information while the colData contain cell-level metadata, including image names. The assay slots contain marker expression per cell values: the counts assay contains average ion counts per cell whereas the other assays contain counts transformation(s) (details available in the documentation of each dataset).

2. Multichannel images formatted into a CytoImageList class object. This object contains multichannel images and metadata, including channel names and image names.

3. Cell segmentation masks formatted into a CytoImageList class object. This object contains single-channel images representing cell segmentation masks and metadata, including image names.

The three data objects can be mapped using the image names contained in the metadata of each object. Details are available in the vignette (see above).

For more information about the SingleCellExperiment and CytoImageList objects, please refer to the SingleCellExperiment and cytomapper packages, respectively.

Available datasets

List of available datasets

• JacksonFischer2020: Tumour tissue from patients with breast cancer.
  • Documentation: JacksonFischer2020Data.
  • Publication: Jackson, Fischer et al. Nature (2020) 578:615–620
• ZanotelliSpheroids2020: 3D spheroids generated from different cell lines.
  • Documentation: ZanotelliSpheroids2020Data.
  • Publication: Zanotelli et al. Mol Syst Biol (2020) 16:e9798.
• DamondPancreas2019: Pancreas sections from donors with type 1 diabetes.
  • Documentation: DamondPancreas2019Data.
  • Publication: Damond et al. Cell Metab (2019) 29(3):755-768.e5.

View available datasets directly in R

In R, currently available datasets can be viewed with:

imc <- imcdatasets::listDatasets()
imc <- as.data.frame(imc)
imc

Detailed information about each dataset is available in the help pages (e.g., ?JacksonFischer2020Data).

Alternately, available datasets can be viewed without installing imcdatasets with ExperimentHub, as following:

# Install and load the ExperimentHub package
if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("ExperimentHub")
library(ExperimentHub)

# View available datasets
eh <- ExperimentHub()
query(eh, "imcdatasets")

For more information, please refer to the ExperimentHub vignette.

Citation

Damond N, Eling N, Fischer J (2022). imcdatasets: Collection of publicly available imaging mass cytometry (IMC) datasets. R package version 1.5.2, https://github.com/BodenmillerGroup/imcdatasets.

Authors

Nicolas Damond (author, maintainer) Nils Eling (contributor) Jana Fischer (contributor)

References

[1] Giesen et al. Nat Methods. 2014. 11(4):417-22