Hi, all - hopefully this is the right forum for this, let me know if I should continue this with smaller/different group.
At Open Force Field, we are working to make sure we can export systems with 4-point water models like OPC using the OpenFF toolkits in a way that they can be read by sander/pmemd/other tools and programs in the Amber ecosystem.
If the "right" way to run rigid 4 point models it is just construct a properly formatted prmtop, we can do that, we just need to make sure we know the developer-approved way, as the specific formats do not be well-documented in the manual or elsewhere on the site. But are there are other places where instructions need to be passed in? Ideally, it would be good to have this all documented in the manual so we can keep things updated if there are changes.
We've done some reverse-engineering to make sure we understand the problem, but don't want to rely on undocumented features if at all possible.
Looking at water boxes for TIP3P and OPC generated by tleap in Amber25, I see that the inpcrd's pretty much just reflect the different number of particles (nwaters*3 for TIP3P, nwaters*4 for OPC). For the prmtops, it's more complicated:
*
OPC waters have EPW for the virtual site ATOM_NAME, and EP as the AMBER_ATOM_TYPE as well as 0 for the virtual sites ATOMIC_NUMBER and 0.0 for MASS, as well as some differences in NONBONDED_PARM_INDEX I am not sure understand.
*
The virtual site is omitted from bonds, which are defined by harmonic bonds (rather than bonds and angles being defined)
*
Some differences in POINTERS that I can't quite parse because the number of fields in the description of the prmtop format on the website is off by one with the one amber25 generated for these systems.
*
BONDS_INC_HYDROGEN and BONDS_WITHOUT_HYDROGEN have some differences I haven' t parsed yet, as do the EXCLUDED_ATOMS_LIST, and TREE_CHAIN_CLASSIFICATION.
Are there other places where the "treat this as a 4-point rigid water molecule with a virtual site" is signaled to the sander/pmemd?
Longer term, it would be good to understand if there are ways to encode other virtual sites in prmtop, but that is probably a question for another time.
Thanks so much!
~~~~~~~~~~~~~~~~
Michael Shirts
michael.shirts.colorado.edu
http://www.colorado.edu/lab/shirtsgroup/
Phone: (303) 735-7860
Office: JSCBB C123
Professor, Department of Chemical and Biological Engineering
University of Colorado Boulder
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed Apr 22 2026 - 13:00:03 PDT
