We need to cleanly distinguish lists of 'cannonical markers' used as evidence for cell type annotation from lists of markers derived from the dataset being annotated. Note that having a large-ish set of highly enriched derived markers is sometimes used as evidence that a cell set corresponds groups cells belonging to a distinct cell type (which may be novel).
The current schema is causing some confusion:
The core schema has 'evidence markers' which is intended as a place for markers used as evidence for a particular annotation
The CAP extension has a field for 'cannonical markers'.
Chloe took 1 to refer to derived marker sets and 2 to refer to cannonical markers used as evidence for annotation. I took 2 to be a field for collecting general knowledge about canonical markers, independent of whether they are expressed in the cell set or used as evidence for cell identity.
Neither of these fields is suitable for derived markers.
We need to cleanly distinguish lists of 'cannonical markers' used as evidence for cell type annotation from lists of markers derived from the dataset being annotated. Note that having a large-ish set of highly enriched derived markers is sometimes used as evidence that a cell set corresponds groups cells belonging to a distinct cell type (which may be novel).
The current schema is causing some confusion:
Chloe took 1 to refer to derived marker sets and 2 to refer to cannonical markers used as evidence for annotation. I took 2 to be a field for collecting general knowledge about canonical markers, independent of whether they are expressed in the cell set or used as evidence for cell identity.
Neither of these fields is suitable for derived markers.
Proposed solution
new fields:
Cannonical marker evidence: { description TBA - drafts on JIRA } Derived markers: { description TBA - drafts on JIRA }