Phenotyping is the process of classifying single cells into cell types, which allows us to better understand their role in each tissue and in cell-cell interactions. It is one of the main reasons to perform IHC and/or IF techniques, in addition to understanding protein expression patterns.
A mapping of different cell lineages and states to biomarker expression for human cells. Different biomarkers are used to distinguish between cell phenotypes.
Cell phenotypes are primarily assigned by the biomarker expression found in each cell. While there may be additional phenotyping that can be done based on cell morphology, this approach has not yet been implemented on our platform. For phenotyping based on biomarker expression, we use two methods: unsupervised clustering and gating.
<aside> 📝 We suggest starting with unsupervised clustering, then labeling the phenotypes of the clusters based on visual assessment of the clusters in the Visualizer using the biomarker expression heatmap as a guide.
If some clusters need to be further stratified into subclusters to distinguish between relevant phenotypes, we typically do so through Workbench.
</aside>
Once you have created a Cell Phenotype Annotation with one of the above extensions you may also manually fine tune the annotations on each cell from within the Visualizer. To learn more, please see: Editing Annotation Classes for Cell Overlay Channels
From either Unsupervised Clustering or Gating, cells will be assigned phenotypes. The phenotypes are added as another column to the table containing all of the cells, and can be viewed as Cell Overlays in the Visualizer.
After evaluation through the Visualizer, any issues can be corrected through the annotation editing tool in the Visualizer. Once the phenotype assignments look correct, then it is time to perform analysis on the Explorer, which allows comparisons between the phenotypes present in different groups of samples (e.g., diseased vs. normal tissue) to be made.
Having the phenotype annotations pushed to the Explorer also opens up Neighborhood Analysis as well as Spatial Neighbor Distance Analysis.