Closed malachig closed 6 years ago
On the licensing issue:
Another option that has been recommended to us several times is NCI Thesaurus. It is released as Public Domain open content license.
HPO appears to have periodic stable releases: http://human-phenotype-ontology.github.io/ http://human-phenotype-ontology.github.io/downloads.html
It might be nice to support both HPO and NCI thesaurus.
We didn't really get a chance to talk about this further during the meeting. Are phenotypes something we want to add evidence items or to assertions? Or to both?
Good question. I think to both
Let's go with HPO for now as well.
I think we decided to require an HPO term to first be added using an HPO ID, pulling that from their API. Only HPO IDs that have been added will be in the type ahead search for text. This reduces the crazy number of options in the type ahead search.
It looks like we could use these to create EIDs for brain metastases: http://compbio.charite.de/hpoweb/showterm?id=HP:0030692
Choose say NSCLC from DOID and then the above from HPO.
When we provide the message that people should identify an HPO term, they should be directed to the following URL. I find the home page non-intuitive and would rather direct them immediately to the term browser that is relevant. http://www.human-phenotype-ontology.org/hpoweb?id=HP:0000118
@arpaddanos I'm not sure I agree with that approach but we should definitely discuss that in our next CIViC meeting.
HPO is now integrated! Great job all. Closing this issue.
It has been suggested that we should allow curation of associated phenotypes in addition to diseases. For example, for many VHL variants the appropriate disease is
Von-Hippel Lindau
but there is value to being to further specific which phenotypes are observed in each patient within this syndrome (e.g.pheochromocytoma
,renal cell carcinoma
,hemangioblastoma of the CNS
, etc.).It has been suggested that we use HPO terms for this purpose similarly to how we use DO terms for the disease.