Tentative set of tools and scripts for analysing spatial transcriptomic data with the resolve platform.
The script is based on this Seurat vignette.
The script currently requires a development version of Seurat from the feat/imaging branch
Assumes the input cell segmentation was generated with: https://codebase.helmholtz.cloud/resolve_tools/resolve-processing
However, it should be adaptable to use segmentation output from other tools.
The ReadResolve function expects a folder with the following files:
*-cell_data.csv: csv file with this header:cell,area,centroid.y,centroid.x,label,GENE1,GENE2,...cell= table index, not used.centroid.y,centroid.x= centroid coordinates (µm or pixel, see theuse.micronargument)area= area in pixel^2GENE1,GENE2,...= transcript counts per cell.
*-filtered_transcripts.txt(either the raw output from resolve or the deduplicated output from MindaGap): csv file with no header and these column:- x: pixels
- y: pixels
- z: not used, but required
- gene name
- quality: not used, optional
*-roi.zip(optional, if not provided the centroids are used), zip format used by FiJi.*_mask.tiff: segmentation mask from cellpose or similar tool (tiff file, used only to get the total width and height, so any image would work)*-gridfilled.tiff: optional image to be added to the Seurat object for visualization. Generally we use the output from MindaGap, but any image would work.