search

SysBioChalmers / Human-GEM

The generic genome-scale metabolic model of Homo sapiens
https://sysbiochalmers.github.io/Human-GEM-guide/
Creative Commons Attribution 4.0 International
98 stars 41 forks source link

Chemical structure incorrect #766

Open Christoff1993 opened 11 months ago

Christoff1993 commented 11 months ago

I have noticed two erroneous structures on the Human Metabolic Atlas website.

The first is the structure indicated on the entry for cis-vaccenic acid on the Human Metabolic Atlas website (https://metabolicatlas.org/explore/Human-GEM/gem-browser/metabolite/MAM01585c) contains a trans double bond. The compound shown looks like vaccenic acid (which is has a trans double bond) and not cis-vaccenic acid. This is the case in all compartments where cis-vaccenic acid is found.

The second is for cis-vaccenoy-CoA, which contains three cis double bonds and rather looks like the structure for a CoA ester of gamma-linolenic acid (https://metabolicatlas.org/explore/Human-GEM/gem-browser/metabolite/MAM01586c).

haowang-bioinfo commented 11 months ago

@Christoff1993 Thanks for reporting these errors

could you please help to link MAM01585 and MAM01586 to their correct structures, such as from ChEBI?

mihai-sysbio commented 11 months ago

For reference, this is how the structure image is obtained on Metabolic Atlas with RDKit:

this.mol = Object.freeze(RDKit.get_mol(this.smiles));

Is this perhaps a case of SMILES confusion?

mihai-sysbio commented 11 months ago

Is this perhaps a case of SMILES confusion?

On second thought, that is unlikely.

It looks like on Metabolic Atlas the SMILES is fetched by mapping the information from the model with GotEnzymes compounds, which includes the needed SMILES.

https://github.com/MetabolicAtlas/MetabolicAtlas/blob/58db8ab6d19eb65942b9d93aac8b11a89a1d91ff/api/src/gotEnzymes/compound.js#L48-L58

Since this issue cannot be resolved in Human-GEM, my suggestion would be to transfer it to Metabolic Atlas.