While annotation table(s) should only expose marker gene names, it would be useful for CAS to support a more complete representation
A set of markers can be represented as a list of objects with keys for gene_name & gene_id. Names supplied in tables can be cross-checked against vars in the linked h5ad file.
It would also be useful for the object support extra metadata about the list:
- if it is derived from the dataset - how was it derived? (Free text may suffice)
- What is the confidence score for identifying cells in linked dataset with this marker gene set (NS-Forest provides this).
Related issue #127
The only compulsory field = name, but we would aim to fill the rest out via a mix of lookup and curation.
While annotation table(s) should only expose marker gene names, it would be useful for CAS to support a more complete representation
A set of markers can be represented as a list of objects with keys for gene_name & gene_id. Names supplied in tables can be cross-checked against vars in the linked h5ad file.
It would also be useful for the object support extra metadata about the list:
Related issue #127
The only compulsory field = name, but we would aim to fill the rest out via a mix of lookup and curation.