Set atom types for residue based on dictionary

[Tristan Croll]

AddH is getting things wrong for the modified nucleobase 1MG in 6hrm (partially because the starting geometry is a little wonky):

• O6 is treated as single-bonded (actually double)
• the ribose is treated like an aromatic ring (to be fair, the geometry *is* somewhat too close to planar, and the chirality is wrong at O2')

In cases like this, it would be great to have the option to add hydrogens strictly according to the CCD definition of the residue.

[tic20@…]

On the other hand, turns out that the *dictionary* definition of 7MG is 
wrong, so that's not a perfect solution either. Sigh...

On 2019-02-14 10:27, ChimeraX wrote:

[tic20@…]

Belay that. 7MG isn't wrong, the authors just chose the wrong residue ID 
(should be MG7).

On 2019-02-14 12:15, ChimeraX wrote:

[Eric Pettersen]

With the base-pairing flipped like that, there is no hydrogen bonding stabilizing the keto form of the guanine, so it may well favor the enol form. I wouldn't classify adding the proton there as "wrong". The C→O bond length is indicative of a single bond.

Looking the "sugar" ring, man, I can see why it got treated as furan. No impartial observer would think those carbons are tetrahedral.

I disfavor using CCD templates for protonation since they use neutral protonation states rather than biologically relevant (pH 7.4) protonation states. The typical remedy here is to explicitly assign the correct IDATM types to the ring atoms before protonation. Nonetheless I could allow using CCD templates as an option. I've added it to my low-priority to-do list.

--Eric

[tic20@…]

From my perspective, the most important thing is to first get *something* consistent with the force field so I can improve the geometry and fit to the map. Protonation states are a secondary issue for me - important, yes, but can only be predicted accurately *after* the geometry’s right.

 
 
Tristan Croll
Research Fellow
Cambridge Institute for Medical Research
University of Cambridge CB2 0XY
 

 

[Eric Pettersen]

What you really want is a tool that sets the atom types based on templates as much as possible, which AddH will then naturally use with no changes. Do you agree?

[tic20@…]

That would make sense, yes.

 

[Eric Pettersen]

Not fixed as per my suggestion (setting atom types based on templating), but nonetheless addressed by the addh "template" keyword option.