Conversation
Stanvk
changed the title
There was a problem hiding this comment.
Is there a reason this separate file is required? We don't have separate YAML files for the other force fields.
There was a problem hiding this comment.
Thank you for considering this PR.
Indeed, this file is not necessary and can be excluded.
There was a problem hiding this comment.
What is the provenance of this file? I can't find it anywhere in the AmberTools25 distribution.
There was a problem hiding this comment.
Using the distributed LEaP input file failed with
parmed.exceptions.ParameterWarning: Some residue templates using unavailable AtomTypes were found. They will not be written to the ffxml as write_unused is set to False.
As a workaround, this file is a copy of the DNA.OL21 LEaP input, adapted for OL24 by substituting the updated frcmod. The resulting input file corresponds to the original leap-ff-nucleic-OL24.in file with parm10=loadamberparams parm10.dat; atom type hybridizations; and PDB name maps from the leaprc.DNA.OL21 file included. diff between the generated XML file and the DNA.OL21.xml shows the expected modifications in the proper dihedrals (and, before "fixing" the improper dihedrals, also the different indexing in the improper dihedrals). The DNA.OL24.xml forcefield file has been validated in OpenMM by a single-point energy comparison against the GROMACS topology files (of a solvated DNA duplex), see https://fch.upol.cz/ff_ol/gmxOL24.php.
Please let me know what is considered best practice here.
There was a problem hiding this comment.
What is the purpose of this file? It looks like it is the actual leaprc.DNA.OL24 from the AmberTools25 distribution, except without source oldff/leaprc.ff14SB, and with an added line parm10 = loadamberparams parm10.dat. I'm also not seeing where this file is used, since it doesn't appear to be read in by any of the other LEaP input files or referenced in the the YAML files.
There was a problem hiding this comment.
This file is a second workaround attempt. When passed as the source argument in the YAML input, convert_amber.py runs without errors but produces no XML output (also not by using epretti/ParmEd@2eac516).
There was a problem hiding this comment.
If you use this patched version of ParmEd instead, does this file still change?
We may want to try to keep it consistent. Very unfortunately, when I looked into this in the past, in addition to some bugs in ParmEd regarding improper ordering, I discovered bugs in LEaP itself; it was impractical for OpenMM to try to be "bug-compatible" with LEaP, and I never got a good answer from the Amber mailing list as to what the expected behavior to try to match would be. Consequently, we gave up on trying to get the "right" ordering of the peripheral atoms in Amber impropers since, in short, there didn't seem to be any authoritative answer for what it should be. Sorry for any inconvenience or confusion.
There was a problem hiding this comment.
Generating the DNA.OL21.xml file using epretti/ParmEd@2eac516 produces an identical file to the DNA.OL21.xml that was present in this repository. For consistency across the converted DNA.OL2x forcefield files, I have converted OL24 using the same ParmEd version. I have included the converted forcefield files in dda096a
2 participants
Dear reviewer,
Please consider the following additions and modifications for merge:
Let me know if any further changes are needed.
[1] Zgarbova, M.; Sponer, J.; Jurecka, P, Refinement of the Sugar Puckering Torsion Potential in the AMBER DNA Force Field. J. Chem. Theory Comput. 2025, https://doi.org/10.1021/acs.jctc.4c01100.