openff-toolkit

openff-toolkit copied to clipboard

Describe the bug

Output

Computing environment (please complete the following information):

• Operating system
• Output of running conda list

Additional context

Could you describe what you're trying to do? OpenFF's SMIRNOFF force fields don't use atom types, instead they assign parameters directly to the chemistry of the input molecules. You can check out our documentation for information on how to use our tools: https://docs.openforcefield.org/en/latest/

Hi, I got this error. It was HXT from maestro but I don't know which atom type it should be in openff. raise ValueError( ValueError: I don't know this Ion or something really went wrong! HT HIS -305 connectivity0

OpenFF does not use atom types, it uses direct chemical perception to assign atom parameters.

It means that residue has to be supported as a whole by OpenFF and to check if the residue is correct, OpenFF compares the SMARTS representation of residue with a pool of supported SMARTS. The error you are getting most probably means, that OpenFF can not recognize the residue you have in your structure.

To be able to help you better understand the origin of your problem, could you please provide some further details about what you are doing exactly? Maestro has multiple options to build the protein. One of the options is to add hydrogens to partially resolved residues (e.g. if from the sequence you know that residue should be aspartic acid but in the structure you only have backbone atoms and the residue is terminal, Maestro will place hydrogens to cap each unfulfilled connectivity. The resulting amino acid is not aspartic acid anymore, and won't be supported by OpenFF, but is supported by Maestro), which cannot be supported by OpenFF. If you cap terminal residues in Maestro, this should be fully supported by OpenFF

Hi,

It is a terminal histidine, I used maestro for protein preparation and trying to use openfe for fep calculations. I think the maestro has decided the -COH group as a terminal and H in that case would get a HXT atom type. Here is the residue ATOM 4382 N HIS A 305 -59.759 9.604 -7.160 1.00 26.41 N
ATOM 4383 CA HIS A 305 -60.859 8.622 -7.160 1.00 38.45 C
ATOM 4384 C HIS A 305 -61.427 8.476 -8.565 1.00 72.10 C
ATOM 4385 O HIS A 305 -61.039 9.099 -9.556 1.00 44.74 O
ATOM 4386 CB HIS A 305 -61.975 9.062 -6.180 1.00 38.16 C
ATOM 4387 CG HIS A 305 -61.646 8.994 -4.705 1.00 40.82 C
ATOM 4388 ND1 HIS A 305 -62.560 9.408 -3.732 1.00 42.47 N
ATOM 4389 CD2 HIS A 305 -60.496 8.559 -4.074 1.00 41.57 C
ATOM 4390 CE1 HIS A 305 -61.938 9.231 -2.576 1.00 41.19 C
ATOM 4391 NE2 HIS A 305 -60.713 8.727 -2.718 1.00 41.29 N
ATOM 4392 HT HIS A 305 -60.007 10.549 -6.899 1.00 26.41 H
ATOM 4393 HA HIS A 305 -60.497 7.641 -6.845 1.00 38.45 H
ATOM 4394 HD1 HIS A 305 -62.235 7.749 -8.644 1.00 72.10 H
ATOM 4395 HB3 HIS A 305 -62.864 8.443 -6.316 1.00 38.16 H
ATOM 4396 HB2 HIS A 305 -62.289 10.086 -6.399 1.00 38.16 H
ATOM 4397 HD2 HIS A 305 -59.566 8.174 -4.464 1.00 41.57 H
ATOM 4398 HE1 HIS A 305 -62.379 9.471 -1.620 1.00 41.19 H
ATOM 4399 HE2 HIS A 305 -60.063 8.527 -1.969 1.00 41.29 H

This is exactly what I am talking about. Maestro caps C atom of terminal histidine with HT hydrogen, but the amino acid (if built completely) should get OH or O$^-$. Try using cap terminii flag in maestro. This should add ACE and NME caps and is supported in OpenFF

I think @pbuslaev is right - The terminal HIS can't be capped with a single H.

Also, this error message ValueError: I don't know this Ion or something really went wrong! HT HIS -305 connectivity0 was confusing me since I've never seen it before. I searched for it on GitHub and it appears to be coming from the OpenFE stack, so you might get better responses asking on their repos.

But OpenFE will almost certainly say the same thing as @pbuslaev - The chain should be terminated with a OH or O-. (and having looked at the PDB spec and the residue definitions, a termninal HIS capped with a single H is not a recognized form of the HIS residue, so Maestro is writing an invalid file here)

Does this help @Lili-Cao?