Closed cmungall closed 3 years ago
So, we'd recommend:
bacteria: annotate to GO:0009274 peptidoglycan-based cell wall
plants: annotate to GO:0009505 plant-type cell wall
fungi: GO:0009277 fungal-type cell wall
archae = ??? GO:0030115 S-layer ? (see PMID:31214995)
Rationale:
This creates ragged annotation whereby some things are are annotated to cell wall and sometimes to more specific such as plant cell wall, e.g.
This in turn leads to issues with enrichment (1) [minor] testing more hypotheses than is necessary, increasing the p-value (2) confusing duplicative results as in the enrichment here:
https://github.com/geneontology/helpdesk/issues/322
Surely in all cases it should be possible to make a more specific annotation. Presumably I know if I am making a plant vs fungal cell wall
Note: for cases like this where the differentia is a taxon we could auto-infer deeper placement via gaferencer. But it seems simpler just to put this back on the annotation provider.