After some discussion about the 'disease caused' extension, we decided it shouldn't need to be provided as an annotation extension in most cases because PHI-base already has a list of diseases for common interactions. The disease can also depend on tissue type, in that the same host-pathogen interaction may cause different diseases when it infects different anatomical locations.
So, there is a plan to (eventually) solve this by displaying the expected disease for an interaction after the user uses at least one tissue type annotation extension.
Canto could be provided with a list of disease names, together with the corresponding taxon IDs for the pathogen and host, and maybe the tissue type too, where available. Then Canto could simply look up the disease in this list once the annotation extension is selected, and display the disease name if it is found – otherwise, we could allow users to enter their own name for the disease, similarly to what we do with strains.
It's unlikely that Canto's existing annotation workflow supports the process I've described above. Canto's database would probably require a schema change to support querying for diseases.
This probably isn't a high priority feature, since it won't be needed until we need to curate novel or unseen diseases that aren't already in PHI-base. In the short term, we could encourage users to mention any unusual disease in a comment on the phenotype annotation instead.
jseager7
commented
3 years ago
After some discussion about the 'disease caused' extension, we decided it shouldn't need to be provided as an annotation extension in most cases because PHI-base already has a list of diseases for common interactions. The disease can also depend on tissue type, in that the same host-pathogen interaction may cause different diseases when it infects different anatomical locations.
So, there is a plan to (eventually) solve this by displaying the expected disease for an interaction after the user uses at least one tissue type annotation extension.
Canto could be provided with a list of disease names, together with the corresponding taxon IDs for the pathogen and host, and maybe the tissue type too, where available. Then Canto could simply look up the disease in this list once the annotation extension is selected, and display the disease name if it is found – otherwise, we could allow users to enter their own name for the disease, similarly to what we do with strains.
It's unlikely that Canto's existing annotation workflow supports the process I've described above. Canto's database would probably require a schema change to support querying for diseases.
This probably isn't a high priority feature, since it won't be needed until we need to curate novel or unseen diseases that aren't already in PHI-base. In the short term, we could encourage users to mention any unusual disease in a comment on the phenotype annotation instead.