-

Wed, 03 Sep 2025 21:22:01 +1200

dalke@toots.nu

Any idea how to get the (#cheminformatics) structures for a #ChEBI ontology term? Like, how do I get all the terpenoids, CHEBI:26873?

https://www.ebi.ac.uk/ols4/ontologies/chebi/classes/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FCHEBI_26873?lang=en says there are 11,119 terpenoids but I see no way to download them.

Trying the web API, https://www.ebi.ac.uk/chebi/backend/api/public/ontology/all_children_in_path/?relation=has_functional_parent&entity=CHEBI%3A26873&three_star_only=false&page=1&size=15&download=false says there are 547 (482 if three_star_only=true).

With 'download=true' I get the data in full (unpaged) data in tsv format, with the same counts.

What am I missing?

show all 12 comments

Wed, 03 Sep 2025 21:59:32 +1200 last edited: Wed, 03 Sep 2025 22:28:12 +1200

Egon Willigh☮gen 🟥

egonw@mastodon.social

@dalke so far, I have been using the SD file in combo with the ontology, but @baoilleach can perhaps point you to the this summer introduced new website

Wed, 03 Sep 2025 22:43:06 +1200

Andrew Dalke

dalke@toots.nu

@egonw @baoilleach I tried that last night. The SDFs from https://ftp.ebi.ac.uk/pub/databases/chebi-2/SDF/ are corrupted. From my message to them:

"Consider what should be the record for CHEBI:11, which starts on line 596 of the complete file.

The first two lines are a blank line then the counts line '46 52 0 0 1 0 0 0 0 0999 V2000'. The counts line is supposed to be on line 4."

Two lines are missing, and "the counts line is missing a leading space, causing the field data to be misaligned."

Wed, 03 Sep 2025 22:46:56 +1200 last edited: Wed, 03 Sep 2025 22:49:38 +1200

Charles Tapley Hoyt

cthoyt@scholar.social

@dalke try using the hierarchical descendants query: https://www.ebi.ac.uk/ols4/api/ontologies/chebi/terms/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FCHEBI_26873/hierarchicalDescendants

The hierarchical children query doesn't work since it only gives direct:
https://www.ebi.ac.uk/ols4/api/ontologies/chebi/terms/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FCHEBI_26873/hierarchicalChildren

Wed, 03 Sep 2025 23:09:16 +1200

Andrew Dalke

dalke@toots.nu

@cthoyt Okay, but that shows 25 records, reports size 1000, and doesn't describe how to page.

I prefer the tsv download option as chemfp knows tsv, not json.

Also, is it part of a 1.0 API which may soon disappear?

Thu, 04 Sep 2025 01:19:50 +1200

Charles Tapley Hoyt

cthoyt@scholar.social

@dalke fyi i edited the toot since i had the wrong API endpoint originally - https://www.ebi.ac.uk/ols4/api/ontologies/chebi/terms/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FCHEBI_26873/hierarchicalDescendants seems to get the right things, but it needs paginating.

I've previously written a wrapper around OLS's API (including dealing with pagination) in https://github.com/cthoyt/ols-client, but it's not comprehensive. If you put an issue there, I will add a function for dealing with this endpoint.

I think you're on your own if you want TSV - but this shouldn't be too bad of a transform

Thu, 04 Sep 2025 01:45:39 +1200

Andrew Dalke

dalke@toots.nu

@cthoyt I figured out affixing "?page=0" ... handled paging. Used that to download the JSON. Couldn't figure out a 'download'/'format', or Accept header alternative representation.

Installed your package. I'll be offline for the next 5 hours though.

Thu, 04 Sep 2025 08:42:02 +1200

Andrew Dalke

dalke@toots.nu

@cthoyt I'm so confused. I extracted a SMILES file from the paged JSON files you mentioned, and compared them to the TSV from the ChEBI2 tools.

% wc -l purine*.{tsv,smi} | head -4
1393 purine.tsv
23 purine_alkaloid.tsv
1701 purine.smi
15 purine_alkaloid.smi

Why the different counts? Why does the JSON allow multiple ids and smiles?

All I want is a small demo dataset for a doc section! I'm going with simple - curl the TSV. I'll have a caution commentary too.

Fri, 05 Sep 2025 17:51:35 +1200

Noel O'Boyle

baoilleach@mstdn.science

@dalke @cthoyt Hi Andrew. Can you email help with any unusual situations? Would be great to get problems ironed out as we need to transition to ChEBI 2 in a compressed timeframe.

Fri, 05 Sep 2025 21:37:01 +1200

Andrew Dalke

dalke@toots.nu

@baoilleach

Your helpdesk confirmed the ChEBI2 SDFs are invalid and said the dev will look into when back from leave.

As for the web APIs, I don't know your data model, I don't know the endpoint @cthoyt pointed to, I don't know if they use the same reference data set - I simply don't know what to expect.

And I haven't even looked for an intro guide - I thought I could just jump in and get some example data using my now 10-years-dated knowledge from last really looking at ChEBI.

Fri, 05 Sep 2025 22:09:28 +1200

Noel O'Boyle

baoilleach@mstdn.science

@dalke @cthoyt even so. Just flag it up.

Sat, 06 Sep 2025 00:10:35 +1200

Andrew Dalke

dalke@toots.nu

@baoilleach @cthoyt

I'm pretty sure it's due to me not understanding the difference between "is_a" and "has_functional_parent".

Sat, 06 Sep 2025 01:51:15 +1200 last edited: Sat, 06 Sep 2025 05:57:12 +1200

Andrew Dalke

dalke@toots.nu

@baoilleach

Just sent a help desk message starting:
====
When I use the the "all_children_in_path" REST API for ChEBI2, the "smiles" column contains a SMILES which is different than in the "<SMILES>" data item in the SDF download.

The following reproducible shows how RDKit does not parse the REST download SMILES but can parse the SMILES data item from the SDF.
====