merops IDs are not resolving to the correct page

[cmungall]

There are two entries for this one database in bioregistry; the split was made by identifiers.org:

• merops.family
• merops.inhibitor

This is split doesn't make any sense as I'll explain below

The example IDs:

https://bioregistry.io/reference/merops.family:S1
https://bioregistry.io/reference/merops.inhibitor:I31.952

Don't resolve to individual pages. They just take me to the top level https://www.ebi.ac.uk/merops/ -- this is presumably because bioregistry has a sanger URL, and the redirect doesn't preserve the query parameters.

I think if merops.sanger.ac.uk ==> ebi.ac.uk/merops/ then these particular entries will be fixed, but there are still major issues

If identifiers.org is going to split merops, I don't think the current division makes sense

MEROPS is actually divided into peptidases and inhibitors. Within each of these categories there are divisions into

• the entities themselves - often corresponding to individual genes/proteins
• families
• clans

Example peptidase classification:

• https://www.ebi.ac.uk/merops/cgi-bin/clansum?clan=MA
  • https://www.ebi.ac.uk/merops/cgi-bin/famsum?family=M12
    • https://www.ebi.ac.uk/merops/cgi-bin/pepsum?id=M12.334

The leaf nodes typically equate to individual genes or proteins, but there are also groupings e.g. https://www.ebi.ac.uk/merops/cgi-bin/pepsum?id=C01.UNA

Example inhibitor classification:

• https://www.ebi.ac.uk/merops/cgi-bin/clansum?clan=IA
  • https://www.ebi.ac.uk/merops/cgi-bin/famsum?family=I1
    • https://www.ebi.ac.uk/merops/cgi-bin/pepsum?id=I01.979

If we want to subdivide the merops namespace in order to support resolution, then we should use clan/family/entry -- the same URLs are used regardless of whether we are talking about peptidases or inhibitors.

E.g.

• merops.clan => https://www.ebi.ac.uk/merops/cgi-bin/clansum?clan=
• merops.family => https://www.ebi.ac.uk/merops/cgi-bin/famsum?family=
• merops.entry => https://www.ebi.ac.uk/merops/cgi-bin/pepsum?id=

I don't have a better term than than "entry" - I don't think "protein" is good as the "pepsum" entries are sometimes themselves subfamilies. We could just go with their own nomenclature of "pep"

Note that if we did split like this, this would be something of a hard fork from identifiers.org, unless we can coordinate with them to make the same change.

IMO the whole decision by identifiers.org to split databases into sub-namespaces is really problematic but I have commented on this elsewhere.

I believe merops isn't updated any more so I'm not sure how high a priority this all is.

FWIW, this is the entry we have for GO:

- database: MEROPS
  name: MEROPS peptidase database
  generic_urls:
    - https://www.ebi.ac.uk/merops/
  entity_types:
    - type_name: protein
      type_id: PR:000000001
      url_syntax: https://www.ebi.ac.uk/merops/cgi-bin/pepsum?id=[example_id]
      example_id: MEROPS:A08.001
      example_url: https://www.ebi.ac.uk/merops/cgi-bin/pepsum?id=A08.001

Note that unlike identifiers.org we didn't partition the namespace into two, it's all one. But this only works for the leaf nodes in merops. We made a different category error from identifiers.org. But using our scheme we would simply add the new types under the general MEROPS entry.

[cthoyt]

Okay so we can take the following steps:

1. Add an entry for merops.clan
2. Collapse merops and merops.inhibitor to a single prefix merops.entry
3. Update URIs for merops.entry and merops.family

As each of these have separate entity types and URI format strings, I still disagree about having a combine prefix for all of them. The only case where this makes sense is if there's also an "omni"-URI format string

[cthoyt]

@cmungall alright problem solved