Tuesday 25 September 2018

How to do a substructure search of ChEMBL (2)

I previously wrote some notes on how to do a substructure search of ChEMBL.

I've been doing some more and found some funny/interesting things.

1) Strange things with aromatic rings!
I wanted to do a search that would hit both menadione (a napthaquinone, image from wikipedia):

Skeltal formula
and also hit anthraquinones like diacerein (image from wikipedia):
Structural formula of diacerein
I first tried this with this substructure but it only hit menadione and not the anthraquinones:










Then after discussing with Noel (thanks Noel!) I found out that I need to specify S/A (single/aromatic) and D/A (double/aromatic) bonds, like this:






This found both the anthraquinones and naphthaquinones, hurray!

Note: I found that the ability to specify S/A and D/A bonds is only available in the current ChEMBL website, not the new ChEMBL beta website, which seems to use a different sketcher from ChemAxon.
Note 2 (1-Oct-2018): I heard back from the ChEMBL website that you can specify S/A bond on the ChEMBL beta website by selecting the bond and pressing 14, and D/A by pressing 24 (and 'any' bond by pressing 0). Thanks to David Mendez from ChEMBL! 

2) Specifying any atom with 'A'
I wanted to do a search which would pick up both

and




I found I could do this with the 'A' (any atom), like this:




Note: on the ChEMBL beta website, I see '*', but if I select this and press 'A'  it turns to A.

3) Wierd things I can't explain

I found that when I do this attached substructure search

 it gets a hit to dipananone https://www.ebi.ac.uk/chembldb/index.php/compound/inspect/CHEMBL2111157 when I do the search using the ChEBML beta interface, but not when I do the exact same substructure search using the normal CHEMBL interface. I'm not sure why this is? I sent an email to the ChEMBL helpdesk, waiting for a reply...

4) Starting from a particular structure
If you have a problem drawing in your structure, you can save a related compound as its mol file from the ChEMBL website, and then click on the 'open file' button within the sketcher on the ChEMBL homepage, and open that mol file to be your starting point for editing.

Acknowledgements
Thanks to Noel O'Blog and John Mayfield for help!
And also a big thanks to David Mendez from ChEMBL!

No comments: