ebi-chebi / ChEBI

Chemical Entities of Biological Interest (ChEBI) is a freely available dictionary of molecular entities focused on ‘small’ chemical compounds.
https://www.ebi.ac.uk/chebi
Creative Commons Attribution 4.0 International
42 stars 10 forks source link

CHEBI:4767 'elastin' has been deleted from CHEBI #3060

Open muthuvenkat opened 9 years ago

muthuvenkat commented 9 years ago

It's OBO policy to never delete an ID - they are always obsoleted or merged.

Can you add an obsoletion record for this? Ideally this would indicate a possible replacement term, and ideally obsoletion of proteins would be coordinated with PRO so that a replacement is available.

Reported by: cmungall

muthuvenkat commented 9 years ago

Hi Chris,

behind the scenes in our database, we do keep track of the deleted IDs. However, our OBO export process does not export them as part of the output file. This is something we could build in for the future but cannot directly support just at the moment.

This was a protein term -- we agreed with the PRO and the OBO community a long time ago that proteins were not in scope for ChEBI and hence we've been removing them ad hoc as we get the chance, which was formally announced at some stage years ago. I thought GO was using PRO for these terms?

Cheers, Janna

Original comment by: jannahastings

muthuvenkat commented 9 years ago

Ok we can do this but historically CHEBI has had a much higher coverage of glycoproteins etc

Original comment by: cmungall

muthuvenkat commented 9 years ago

However, our OBO export process does not export them as part of the output file. This is something we could build in for the future but cannot directly support just at the moment.

I think this should be a priority if at all possible. Do you also keep records of reasons for obsoletion and potential replacement terms?

Original comment by: djs93

cmungall commented 6 years ago

Is there any update to this.

Another thing I note is that CHEBI seems to sometimes merge one class into another (alt_id in obo format) even when the meaning is different.

The example I have at hand is CHEBI:26619 which used to be the ID for 'secondary metabolite' (this is going back a few years). This was merged into CHEBI:25212 ('metabolite'). This means that in GO, terms concerning secondary metabolism were axiomatized using CHEBI:26619. We generally assume OBO policy and that merges can be repaired automatically (obsoletions with no 'replaced by' tag are not), but this renders the axiom incorrect. It would be good if CHEBI could make distinct obsolete classes, with a 'consider' link when the meaning is different.