IMMUcan panel 1 cell type classification

This repo contains Imaging Mass Cytometry data from 179 cancer patients collected during the IHI2 funded IMMUcan project.

Background:

IMMUcan (https://immucan.eu/) collects samples from cancer patients for multimodal analysis. Amongst those Imaging Mass Cytometry (IMC). Samples collection occurs over 6 years. For reproducible cell typing over time, a cell phenotype classifier was built from manually annotated data. The classifier was trained on a first batch of annotated images (V1) and then trained again with a second round of annotated images (V3). A final classifier was build which is used within IMMUcan to classify cell phenotypes for IMC panel 1 for all cancer patients.

The dataset contains:

sce_labelled_V1.rds_ is the `SingleCellExperiment` object containing all labelled cells from the first batch of labelling

sce_labelled_V3.rds_ is the `SingleCellExperiment` object containing all labelled cells from the second batch of labelling

rf_images_DCfix.rds_ is the RF model which can be used for cell type predictions

sce_raw.rds_ is the `SingleCellExperiment` object containing all cells from the 179 images. the trained classifier has been applied to this dataset which can be used to view the classification results of cell phenotypes on images using the bioconductor packages cytomapper or cytoviewer. 

all_images.rds a `CytoImageList` object which contains all the images from all images

all_masks.rds a `CytoImageList` object which contains all single-cell segmentation masks from all images

a zip archive `images_masks` containing the tiff images and masks from all 179 images, a CytoImageList Object for images (all_images.rds) and a CytoListImage object for all masks (all_masks.rds)