How should unmapped elements be indicated?

[cmungall]

I frequently want to be able to state: I have looked in object_source for a matching or close concept to subject_id but am confident there is nothing.

How should I state this?

Can we make object_id/name nullable (where we have a convention that '' in the TSV is null)?

alternatively we could have a CURIE/URI for NoMapping, analogous to owl:Thing.

@wdduncan we need this for the nmdc sssom

I'm also using a pseudo-sssom for cob-to-external where I have blank entries for not mapped

[wdduncan]

I like the idea of NoMapping, or perhaps ssssom:NoMapping.

[matentzn]

My main worry is that this statement is a statement on mapping level that requires for its interpretation the context of the mapping (mapping set). My main use case for SSSOM requires the ability to mix and match mappings, which sort of decontextualises them. I guess we could say that the interpretation of sssom:NoMapping is "no mapping in source", and require the object_source to be specified- when sssom:NoMapping is used. We could also use sssom:Nothing instead of noMappig to retain a bit the "object-ness" of the terminology in the object_id column.. I dont mind either way. Something like:

subject_id	relation_id	object_id	match_type	object_source
HPO:001	owl:equivalentTo	sssom:Nothing	sssom:HandCurated	mp.owl

[wdduncan]

We could also use sssom:Nothing instead of noMappig to retain a bit the "object-ness" This makes sense to me!

Also, you could include a column that for whether a mapping exists. "1" means a mapping does, "0" means a mapping does not. Personally, I prefer sssom:Nothing.

[mellybelly]

Needs to be clear that a mapping was sought and not found, regardless of implementation.

[cmungall]

makes sense

On Tue, Aug 18, 2020 at 4:34 AM Nico Matentzoglu [email protected] wrote:

My main worry is that this statement is a statement on mapping level that requires for its interpretation the context of the mapping (mapping set). My main use case for SSSOM requires the ability to mix and match mappings, which sort of decontextualises them. I guess we could say that the interpretation of sssom:NoMapping is "no mapping in source", and require the object_source to be specified- when sssom:NoMapping is used. We could also use sssom:Nothing instead of noMappig to retain a bit the "object-ness" of the terminology in the object_id column.. I dont mind either way. Something like: subject_id relation_id object_id match_type object_source HPO:001 owl:equivalentTo sssom:Nothing sssom:HandCurated mp.owl

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/OBOFoundry/SSSOM/issues/28#issuecomment-675425389, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMMOPWHDMFMU7M3R5I6ELSBJRK3ANCNFSM4QCPWJWQ .

[cmungall]

type=curated & object_id=sssom: Nothing -> a human was not able to find a mapping type=lexical & object_id=sssom: Nothing -> lexical methods not able to find a mapping

sufficient?

On Wed, Aug 19, 2020 at 10:49 AM Melissa Haendel [email protected] wrote:

Needs to be clear that a mapping was sought and not found, regardless of implementation.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/OBOFoundry/SSSOM/issues/28#issuecomment-676570062, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMMOLLGEJFZHGQIJ5HFI3SBQGBHANCNFSM4QCPWJWQ .

[cmungall]

Maybe rather than sssom:Nothing, something like sssom:NoTermFound (to avoid confusing with owl:Nothing)

[wdduncan]

Do we want a special predicate for such cases? E.g.: sssom:noMatch

[cmungall]

we want to be able to say no exact matches, no close matches etc

On Tue, Oct 13, 2020 at 1:28 PM Bill Duncan [email protected] wrote:

Do we want a special predicate for such cases? E.g.: sssom:noMatch

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/OBOFoundry/SSSOM/issues/28#issuecomment-707990393, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMMOKHKWSWBVYHKKKFXPTSKS2ANANCNFSM4QCPWJWQ .

[matentzn]

This tickets discusses three things:

1. How to declare that there is no mapping of a kind
2. How to declare that a mapping is false (negated mapping)
3. How to capture that the above were determined

For 1, the going proposal is:

subject_id	relation_id	object_id	match_type	object_source
HP:001	owl:equivalentTo	sssom:NoTermFound	sssom:HumanCuration	mp.owl

to indicate that given the term HP:001, there was no term in the object_source that a human curator could find would map. Are we all in agreement of that?

For 2 EDIT: moved discussion to #40

[matentzn]

The negated mapping question is discussed in #40 so this ticket is now purely about the question how to specify unmapped elements.

[matentzn]

Vote: Proposal for modelling unmapped elements.

subject_id	relation_id	object_id	match_type	object_source
HP:001	owl:equivalentTo	sssom:NoTermFound	sssom:HumanCuration	mp.owl

To say that: "According to a human curator, there was no term in object_source that could be mapped to HP:001 using the owl:equivalentTo predicate".

• 👍: Agree with proposal
• 👎: Do not agree with proposal (see comment for alternative idea).