Add first version of first set of MeSH mappings

[rays22]

This PR intends to add the first set of OBO Ontology (CL, UBERON, GO, OBA) term mappings to a set of MeSH terms.

[jamesamcl]

Thanks @rays22 -

My initial comments:

• We should be using skos:exactMatch as with other sssom mapping sets rather than owl:equivalentClass
• Labels are missing from the tsv which makes this difficult to check
• Yaml header is missing from the tsv
• Other metadata also missing; e.g. were these all manual curation?
• Related to above, where do the confidence scores come from?

[jamesamcl]

@matentzn would welcome your feedback too

[rays22]

• We should be using skos:exactMatch as with other sssom mapping sets rather than owl:equivalentClass
• Labels are missing from the tsv which makes this difficult to check
• Other metadata also missing; e.g. were these all manual curation?
• Related to above, where do the confidence scores come from?

@jamesamcl ,
The mp_hp_example* files that use owl:equivalentClass had been automatically created by the mapping-commons github template when I created the repo. I will remove them.

I will investigate this:

• Yaml header is missing from the tsv

[rays22]

• When you create mapping registries, please make sure they are registered here: https://github.com/mapping-commons/mapping-commons.github.io/blob/main/mapping-server.yml

[matentzn]

#2

[cthoyt]

FYI we already did a lot of mesh to OBO mappings in Biomappings (https://github.com/biopragmatics/biomappings) - it would be a shame not to reuse those

[jamesamcl]

FYI we already did a lot of mesh to OBO mappings in Biomappings (https://github.com/biopragmatics/biomappings) - it would be a shame not to reuse those

Agreed! Should we mirror them here as SSSOM?

[cthoyt]

Why mirror them? They have a PURL (https://w3id.org/biopragmatics/biomappings/sssom/biomappings.sssom.tsv) for anyone who wants to use them. I think it's a bit misguided to use GitHub as a place for mirroring data (is that what's going on with all of the repos in this org? I disagree with this)

[cthoyt]

the mapping_date seems fake, these likely weren't all manually mapped on this date.

[rays22]

Yes, my friend, these were all manually mapped on 2025-03-26.

[cthoyt]

wow! I am more than impressed. do you have a time estimate on how long this took?

[rays22]

@cthoyt ,

Thank you for your appreciation of my work.

Here is my curation process:

1. I collected the textual definitions of the targeted list of MeSH Concepts, e.g. https://id.nlm.nih.gov/mesh/M0001044.html
2. I created several draft iterations of the term mappings using OLS and the MeSH Concept labels together with their definitions.
3. I double-checked and finalised the mappings for a set of 1,086 terms on 2025-03-26, believed to be correct at that time.

Step 2 took several days to weeks. Nevertheless, the final version of the manual mappings was created on 2025-03-26 (hence the value of mapping_date = 2025-03-26).