A state-of-the-art Leica confocal microscope and careful panel design allow us to image up to 15 antibody-fluorophore combinations in one imaging run.
Virtual cell segmentation
Individual cells are segmented, allowing for the export of the mean fluorescence intensity (MFI), position, and size of each cell for downstream analysis.
Hierarchical gating
We then perform hierarchical gating using the MFI data to determine which cell types and phenotypes are present within the tissue.
Digital cellular map
The position of these cells are then plotted in space, creating a digital cellular map.
Machine-learning-based region classification
We then divide tissues into smaller neighborhoods, which are raster scanned across our dataset. Next, use a machine learning algorithm to cluster these neighborhoods into different regions, based on cellular composition.
Advanced analytics
We then use a suite of analysis tools contained within CytoMAP (Stoltzfus et al. Cell Reports 2020) to interrogate spatial relationships within and between regions. This allows for the comparison of both local-microenvironments and tissue-wide structural features between different groups.
