A guide to IMC data analysis
This website collects tools for imaging mass cytometry™ (IMC™) data analysis. Common analysis steps include image processing, image visualization, and downstream analysis.
Introduction
Imaging mass cytometry is a highly-multiplexed imaging technology that measures the expression of over 40 proteins and RNA molecules (also referred to as markers) in tissues1. For more information on imaging mass cytometry, please refer to the more extended Introduction.
Image processing
Upon image acquisition, raw data are stored in proprietary MCD™ format. This data file contains experiment-specific metadata, one or more multi-channel images (one per acquisition), and the generated panoramas (see Introduction).
The first step of the pre-processing pipeline includes the conversion of images from .mcd to .ome.tiff or .tiff format, which can be read by most image analysis software.
Cell segmentation is performed using random forest or deep learning-based approaches outlined in Image processing. After cell segmentation, intensity features representing marker abundances and morphological features are extracted and exported.
These IMC image processing approaches are implemented in the IMC Segmentation Pipeline and in the steinbock framework.
For more details see Image processing.
Interactive image visualization
Raw .mcd files can be interactively visualized using the MCD Viewer software distributed by Fluidigm®.
The histoCAT toolbox supports interactive image and data visualization, clustering, and spatial analysis of the data generated by IMC. The histoCAT-web web application supports browser-based analysis directly on raw IMC data.
The napari-imc plugin for napari allows joint visualization of multichannel IMC images and panoramas in a shared coordinate system.
A more detailed description of these viewers are described at Image visualization.
Downstream single-cell data analysis
The extracted single-cell features can be read into R using the imcRtools package.
Spatially annotated single-cell data is stored in a SingleCellExperiment or SpatialExperiment container.
The book Orchestrating Single-Cell Analysis with Bioconductor is an excellent resource regarding single-cell data analysis including dimensionality reduction, clustering, and data visualization based on the SingleCellExperiment container.
As part of the Bioconductor project, the imcRtools package supports a variety of spatial data analysis approaches.
For Python-based analysis of single-cell data, steinbock supports the export to anndata format.
For an overview on common analysis steps, please refer to Downstream analysis.
Contributors
Publication
-
Giesen C. et al. (2014) Highly multiplexed imaging of tumor tissues with subcellular resolution by mass cytometry. Nat. Methods, 11, 417–422. ↩