How do I convert LINCS/CMap IDs to ChEMBL/PubChem IDs?

Question

[enricoferrero]

I have some IDs from LINCS/CMap (e.g.: BRD-A00267231) that I need to convert to either ChEMBL or PubChem IDs.

I usually use UniChem for this kind of stuff but BRD identifiers are not recognised. BRD identifiers are used in most (all?) LINCS resources such as clue.io, Harmonizome, L1000CDS2 and LIFE.

Any idea on how I can programmatically convert these IDs?

Thanks!

Answer 1

[Lars Juhl Jensen]

I'm afraid the answer is "with great difficulty, if at all". People in my group tried doing that, and the only way they found was to map via canonical SMILES.

[enricoferrero]

And yet on LIFE I can see BRD to PubChem IDs mappings so it must be possible to do that (without having to go through SMILES):

[Lars Juhl Jensen]

It is quite possible that they did the mapping via SMILES too, though. That is a pretty common approach when combining multiple sources of chemical compounds.

[enricoferrero]

Can you please point me to the method you used to map these IDs to SMILES? Thanks

[Lars Juhl Jensen]

There is sadly no code that I can share with you. What we did was to download structure files associated with each of the two sets of identifiers (e.g. .sdf files). We then used Open Babel to construct canonical SMILES for all the compounds from both resources and compared the canonical SMILES between the two databases to construct an identifier mapping.