Tautomerization of library compounds

Started by rosie4loop

rosie4loop Lv 1

I've noticed some imino form of exocyclic nitrogen in heterocyclic aromatics from library hits.
Example includes one of the hits from puzzle 2313.
(See the second figure in https://fold.it/forum/suggestions/add-a-checkbox-to-allow-filtering-library-compounds-with-bad-groups)

May not be a big problem in the final results of small-molecule design puzzles, since I'd assume some kind of standardization and selection of lower energy tautomers when post-processing players designs.

However in the design process it affect donor-acceptor position and torsions. It'd also affect the protonation estimation in Foldit, assuming generation and selection of 3D isomers/tautomers from downloaded SMILES is done before the protonation step.

It's rather a complicated problem, naturally both forms could exist at a certain ratio, with the dominant form depends on other parts of the structure and the environment.

So I'm curious, how Foldit decide the tautomer to be generated and displayed in the library panel? Some simple rule-based method or pre-calculation of ligand energy?

rmoretti Staff Lv 1

It's arbitrary. We do not currently do any tautomer enumeration/fixing. I believe in large part which one you get is due to whatever the compound library search comes back with. (Which tautomer is annotated in the result.) But even with manual building you can run into issues, with the compound construction picking one tautomer versus another arbitrarily.

Having more fine grained control of tautomers (or allowing players to control the tautomer) is something we should add.