Closed shawntanzk closed 2 years ago
shawntanzk
commented
2 years ago PATO seems to use weird purls for a whole bunch of relations, this should be replaced by RO terms Will import terms into RO and change the IRIs to point to RO purls instead
shawntanzk
commented
2 years ago mappings:
towards > RO:0002503 correlates_with > RO:0002610
decreased_in_magnitude_relative_to > These bunch have no RO term, might have to also look individual case basis? different_in_magnitude_relative_to > perhaps using RO:0002424 (differs in) similar_in_magnitude_relative_to > increased_in_magnitude_relative_to > has_relative_magnitude >
directly_associated_with > based on magnitude too, not in RO, but it doesnt seem to be used in PATO atm? has_cross_section > don't think theres a RO term that describes this well has_dividend_entity > not sure what this is and it isnt used in PATO has_dividend_quality > not sure what this is and it isnt used in PATO has_divisor_entity > not sure what this is and it isnt used in PATO has_divisor_quality > not sure what this is and it isnt used in PATO has_ratio_quality > not sure what this is and it isnt used in PATO inversely_associated_with > seems to have to do with magnitude, not sure there's a good mapping is_magnitude_of > not sure what this is and it isnt used in PATO is_measurement_of > not sure what this is and it isnt used in PATO is_unit_of > not sure what this is and it isnt used in PATO realized_by > not sure what this is and it isnt used in PATO
reciprocal_of > technically sounds like RO:HOM000060 (in apparent orthology relationship with) but the RO term is specific to species homology
singly_occurring_form_of > seems a bit specific to PATO?
shawntanzk
commented
2 years ago @cmungall I noticed that a lot of these terms were added by you - was wondering if these should move to RO or there's a reason to keep them specific to PATO?
anitacaron
commented
2 years ago Nico will contact Chris on Slack.
matentzn
commented
2 years ago Did. pato slack on obo-community
cmungall
commented
2 years ago If they are useful, move them to RO
Some are used in TO, we need to align TO patterns with upheno
shawntanzk
commented
2 years ago If they are useful, move them to RO
@cmungall does useful = actually used in PATO or like generally useful? Happy to try to bit by bit sort this out and migrate to RO
matentzn
commented
2 years ago This is a great question @shawntanzk
I would do this:
Check on OLS whether a relation is used somewhere:
For example, has_divisor_entity is used by FLOPO, MS and PATO:
Curate your list above to look like this:
Then in our tech goal, we can decide to obsolete some relations, or move them to RO.
matentzn
commented
2 years ago And dont spend much time, remember: open ontology work is super exciting and important, but it is a time suck: better we make tiny constant progress then to have someone like @dosumis pull the plug because we get to carried away fixing things that are only tangentially related to our respective projects.
shawntanzk
commented
2 years ago Some of these already exist!
@balhoff should we introduce numerical IDs for these in your opinion?
EDIT by @matentzn: removed CMPO, FLOPO and MS, as they import PATO as a whole
anitacaron
commented
2 years ago @matentzn will discuss with Chris
matentzn
commented
2 years ago Usage API
SPLIT into to
1) Useful in PATO for modeling 2) Design patterns for PATO 3) Some of them are used in general design patterns (plant people xyz) -
finish later
matentzn
commented
2 years ago I have updated @shawntanzk comment above with the proposed actions.
shawntanzk
commented
2 years ago http://purl.obolibrary.org/obo/pato#correlates_with - this is currently being used, but also, it can be mapped to a RO term (https://www.ebi.ac.uk/ols/ontologies/ro/properties?iri=http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FRO_0002610) - will map instead
shawntanzk
commented
2 years ago For realized by, I have decided to deprecate it and leave a comment and term replaced by (it is current not being used and the RO term is not exact)
balhoff
commented
2 years ago @shawntanzk to answer your question above, I do think numerical IDs for those properties would be good.
shawntanzk
commented
2 years ago @shawntanzk to answer your question above, I do think numerical IDs for those properties would be good.
@balhoff - I guess the question then is if they belong in RO? or should we do it like the old CP terms in CL and convert them to PATO IRIs?
matentzn
commented
2 years ago Every OP in RO!
shawntanzk
commented
2 years ago Mapping to new RO terms (will map when they are out in RO): http://purl.obolibrary.org/obo/pato#decreased_in_magnitude_relative_to – RO:0015008 http://purl.obolibrary.org/obo/pato#different_in_magnitude_relative_to – RO:0015006 http://purl.obolibrary.org/obo/pato#similar_in_magnitude_relative_to – RO:0015009 http://purl.obolibrary.org/obo/pato#increased_in_magnitude_relative_to – RO:0015007 http://purl.obolibrary.org/obo/pato#has_relative_magnitude – RO:0015010 http://purl.obolibrary.org/obo/pato#has_cross_section – RO:0015011 http://purl.obolibrary.org/obo/pato#reciprocal_of – RO:0015012
LarsVogt
commented
10 months ago Hi, we have a case in a material science project where we want to be able to model proportions/ratios by relating the dividend value to the divisor value (e.g. density as the ratio of mass over volume). I used pato:hasDividendQuality and pato:hasDivisorQuality for this before but it now neither in PATO nor in RO. Any chance to get them back or an adequate replacement?
matentzn
commented
10 months ago Absolutely, can you make a new relation request in RO for these relations, link this issue and ask @anitacaron for help adding it?
PATO currently probably uses non-RO terms for relations when they should be RO (from the report): see http://dashboard.obofoundry.org/dashboard/pato/fp7.html This needs to be fixed by replacing them with RO terms @shawntanzk to investigate
EDIT: Also, Object Properties should not be using mnemonic names (like RO:towards), see also https://github.com/pato-ontology/pato/issues/355, but real ids like RO:001