search

tgbugs / ontquery

a framework querying ontology terms
MIT License
3 stars 3 forks source link

cell type inconsistent prefix #7

Open bendichter opened 6 years ago

bendichter commented 6 years ago

I was testing out cell type queries and I noticed that there were a bunch of different systems being used. For query('pyramidal cell') it looked like the best result was 'SAO:862606388', for query('mossy cell') the only result was 'NIFEXT:164', and for query('basket cell') the only result was 'CL:0000118'. Will it be a problem that each of these results have a different prefix?

tgbugs commented 6 years ago

In general having different prefixes is not an issue. In particular it should also not be an issue since it is possible to use the 'cell' category, however it appears that category= queries are broken at the moment, which was an unknown issue that I am looking into.

We are in the middle of completely reworking our cell types under a new schema that will be more consistent (and appear under a single prefix as a result). There is an IPython notebook with some of the details for that which is now a bit out of date since it needs a section on how to use OntTerm and OntQuery.

bendichter commented 6 years ago

OK good. One of the main purposes for NWB to use ontologies is to facilitate consistency. It might be a problem for consistency if multiple suitable terms come up for a query like query('pyramidal cell'). We can get around this if we have conventions like "always use NIF for cell type," however it seems a rule like that could not currently be implemented because there are some cell types (e.g. basket cell) that are supported by other prefixes but not NIF, and there is no prefix that supports them all. I'm glad to hear you are already working on a project to rework cell types which should address this issue.

tgbugs commented 6 years ago

Category issue tracked at https://github.com/SciGraph/SciGraph/issues/263.