In the Ontology Quality Assessment, I found that there was an unknown prefix oma (see https://cthoyt.com/oquat/unknowns/prefix/oma). This apparently only appears in definition xrefs in PRO entries that shadow UniProt identifiers - it's not really an appropriate use of this field, and even further, it just duplicates the identifier already available in the term's URI itself as well as the dbxref to uniprot.

For example PR pages that shadow UniProt identifiers like https://proconsortium.org/app/entry/PR:Q01112 have links to OMA as if they're xrefs.

Related: I also added OMA as a provider in the Bioregistry in biopragmatics/bioregistry#433