nf-core pipeline · cytometry, image-analysis, image-processing
nf-core/imcyto
A Nextflow pipeline for image segmentation and single-cell expression extraction from Imaging Mass Cytometry data.
What nf-core/imcyto is
nf-core/imcyto is a bioinformatics analysis pipeline designed for image segmentation and extracting single-cell expression data. While engineered primarily for Imaging Mass Cytometry (IMC) experiments, it is adaptable to other imaging techniques, including immunofluorescence and immunohistochemistry.
Originally developed by the Bioinformatics & Biostatistics Group at The Francis Crick Institute in collaboration with multiple researchers, the workflow also incorporates methods and scripts established by the Bodenmiller Lab. It standardises the transition from raw multi-channel acquisition files to structured single-cell measurements.
High-plex tissue imaging often yields complex multi-channel images requiring robust segmentation masks to differentiate individual cells accurately. By packaging these preprocessing, pixel classification, and segmentation tasks into a reproducible pipeline, nf-core/imcyto enables automated derivation of quantitative single-cell data across regions of interest.
What the pipeline does
The pipeline starts by reading raw acquisition files (in MCD, OME-TIFF, or TXT format) using imctools. It splits them by region of interest (ROI) and extracts individual TIFF files according to channel definitions supplied in a metadata CSV. Two separate image stacks are generated: a full stack for single-cell expression measurement and an Ilastik stack for mask construction.
CellProfiler applies pre-processing filters to the full stack TIFFs and combines selected channels from the Ilastik stack into a composite RGB image outlining plasma membranes and nuclei. Next, Ilastik performs pixel classification to generate a probabilities map for background, membranes, and nuclei; this stage can optionally be skipped using the --skip_ilastik parameter to proceed directly with the composite TIFF. Finally, CellProfiler uses the probability or composite map along with the filtered full stack to generate a single-cell segmentation mask, overlays this mask onto the full stack, and exports expression data as CSV files.
Key facts
- Pipeline framework
- Nextflow
- Input formats
- .mcd, .ome.tiff, .txt, metadata.csv, .cppipe, .ilp
- Output formats
- .tiff, .csv
- Core dependencies
- imctools, CellProfiler, Ilastik
- Release version
- 1.0.0
- Original development
- The Francis Crick Institute
What you can use it for
- Imaging Mass Cytometry analysis
- Extract individual channel TIFFs from multi-channel IMC acquisition formats and calculate single-cell signal intensities for all detected markers.
- Multi-channel fluorescence or immunohistochemistry segmentation
- Segment cells and extract per-cell expression profiles from standard multiplexed immunofluorescence or immunohistochemistry images.
- Machine-learning assisted cell segmentation
- Classify cellular compartments such as membranes, nuclei, and background via pre-trained Ilastik project files to build single-cell masks.
Run it here
The research agent runs nf-core/imcyto with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
Strengths and limits
Good at
- Supports multiple raw input acquisition formats including MCD, OME-TIFF, and TXT files.
- Allows users to skip the Ilastik pixel classification step using --skip_ilastik if CellProfiler pipelines alone provide sufficient segmentation.
- Separates the analysis channels into dedicated stacks for segmentation masking and single-cell quantification.
Watch out for
- Requires pre-configured CellProfiler pipeline files (.cppipe) and an Ilastik training project file (.ilp) unless pixel classification is skipped.
- Channel names in the input files must strictly correspond to entries supplied in a user-provided metadata CSV file.
Questions
Can I run nf-core/imcyto without my own cluster?
The research agent runs nf-core/imcyto with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
What licence is nf-core/imcyto under?
MIT, like every nf-core pipeline.
Can the pipeline process images other than Imaging Mass Cytometry data?
Yes. Although created for Imaging Mass Cytometry, it is flexible enough to handle other imaging formats such as immunofluorescence and immunohistochemistry.
Is Ilastik strictly required to run the pipeline?
No. If CellProfiler modules alone can reliably segment your images, you can provide the --skip_ilastik flag to bypass Ilastik and use the composite TIFF directly for downstream segmentation.
What inputs are necessary to initiate an analysis run?
You need input image files (.mcd, .ome.tiff, or .txt), a metadata CSV mapping the channels, CellProfiler pipeline files (.cppipe) for full-stack preprocessing, Ilastik stack preprocessing, and segmentation, as well as an Ilastik training parameter file (.ilp) if pixel classification is used.
How to cite it
Febe van Maldegem, Karishma Valand, Megan Cole, Harshil Patel, Mihaela Angelova, Sareena Rana, Emma Colliver, Katey Enfield, Nourdine Bah, Gavin Kelly, Victoria Siu Kwan Tsang, Edurne Mugarza, Christopher Moore, Philip Hobson, Dina Levi, Miriam Molina, Charles Swanton & Julian Downward. Characterisation of tumour microenvironment remodelling following oncogene inhibition in preclinical studies with imaging mass cytometry. Nat Commun. 2021 Oct 8;12(1):5906. doi: 10.1038/s41467-021-26214-x
Written from nf-core/imcyto's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.