Interpreting the output of Predict Cell Types
Predict Cell Types outputs:
- A Cell Clusters () element containing two categories:
- "Cell type (all)" contains the predicted cell types, for each cell in the input matrix.
- "Cell type (high confidence)": if cell type refinement is not performed, it contains the same predicted cell types as "Cell type (all)", but with predictions with probability below 0.5 being replaced with "Unknown". Otherwise, see Cell type refinement. High confidence cell types can be useful for detecting novel cell types.
- Optionally, a Cell Annotations () element with the probabilities assigned for a subset of relevant cell types, for each cell in the input matrix. A cell type is considered relevant if:
- The cell type is the predicted cell type for at least one cell in the "Cell type (all)" category, or
- There is at least one cell with a probability of at least for the cell type.
- Refined (high confidence): "Yes" if "Cell type (high confidence)" was changed through refinement, "No" otherwise.
- Dominant cell type % (high confidence): The percentage of cells in the cluster with the dominant high confidence cell type before refinement.
- Dominant parent cell type % (high confidence): The percentage of cells in the cluster with the dominant parent cell type as parent or ancestor, with high confidence, before refinement.
- Refined (all): "Yes" if "Cell type (all)" was changed through refinement, "No" otherwise.
- Dominant cell type % (all): The percentage of cells in the cluster with the dominant cell type before refinement.
- Dominant parent cell type % (all): The percentage of cells in the cluster with the dominant parent cell type as parent or ancestor before refinement.
Using the outputs, the cells can be colored in a Dimensionality Reduction Plot (see Dimensionality reduction):
- Using the Cell Clusters: by the predicted cell type.
- Using the Cell Annotations: by the probability of having a specific cell type or by whether the initial predicted cell type has been changed due to refinement.
`Dominant cell type % (high confidence)' and `Dominant cell type % (all)' can be used for determining a more appropriate `Minimum dominant cell type in cluster (%)'.
The predicted cell types can be manually further refined in the Dimensionality Reduction Plot (see Manual annotation).
For details on how cell types are predicted, see SVMs for cell type classification.