CMAP indexing error

[BenTanYK]

I am using BioSimSpace 2026.1.0.dev with ambertools 24.8 installed in the same conda environment.

I generate .prm7/.rst7 input files in tleap (see the attached files) and then read these in as a system object:

system = BSS.IO.readMolecules(['input.prm7', 'input.rst7'])

I then try to perform an energy minimisation using OpenMM:

protocol = BSS.Protocol.Minimisation(steps=10000)
process = BSS.Process.OpenMM(system, protocol=protocol, platform='CUDA')

And I get the following error:

I get a similar error when attempting to run OpenMM MD protocols.

Inputs and code example:
files.zip

Any help or suggestions would be much appreciated,
Thanks!

[lohedges]

I've transferred this over to Sire since it appears to be an issue with the parsing of CMAP terms. (At least, how they are processed on write.) To get more info you can use the verbose mode in BioSimSpace, e.g. BSS.setVerbose(True), or just load and save with Sire directly:

In [1]: import sire as sr

In [2]: mols = sr.load("input*7")

In [3]: sr.save(mols, "test.rst7")
Out[3]: ['/home/lester/Code/openbiosim/biosimspace/ben_issue/test.rst7']

In [4]: sr.save(mols, "test.prm7")
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ in <module>:1                                                                                    │
│                                                                                                  │
│ /home/lester/Code/openbiosim/sire/.pixi/envs/dev/lib/python3.12/site-packages/sire/_load.py:687  │
│ in save                                                                                          │
│                                                                                                  │
│   684 │                                                                                          │
│   685 │   molecules = _to_legacy_system(molecules)                                               │
│   686 │                                                                                          │
│ ❱ 687 │   return MoleculeParser.save(molecules, filename, map=map)                               │
│   688                                                                                            │
│   689                                                                                            │
│   690 def load_test_files(files: _Union[_List[str], str], *args, map=None):                      │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
OSError: SireError::io_error: Cannot write the (perhaps some of the ) files as the following errors occurred:
Failed to write the file '/home/lester/Code/openbiosim/biosimspace/ben_issue/test.prm7' using the parser for fileformat
'PRM7'. Errors reported by the parser are:
Something went wrong as there is a cmap between more than one different molecule! (0, 0, 0, 1, 1) (call
sire.error.get_last_error_details() for more info)

I'm not sure if this is a bug in the CMAP parser itself, or an issue with the way the terms are defined in the file. I'm surprised that the issue isn't also raised on load, but perhaps the atoms haven't been split into separate molecules at that stage.

You mention that the input files were generated by tLEaP. Was this via BioSimSpace, i.e. using BSS.Parameters, or did you manually parameterise? If the former, are you able to share the input files and the force field used so that I can try to figure out what's gone wrong, if the latter could you share more details regarding your tLEaP commands or script?

I'll try to track down the specific term(s) that are raising the issue to see if that sheds any more light on things.

[lohedges]

This is the start of the CMAP record section in the topology file:

%FLAG CMAP_COUNT
%FORMAT(2I8)
     380      14
%FLAG CMAP_RESOLUTION
%FORMAT(20I4)
  24  24  24  24  24  24  24  24  24  24  24  24  24  24
%FLAG CMAP_PARAMETER_01
%COMMENT  Gly CMAP
%FORMAT(8F9.5)

Note that it has a comment stating Gly CMAP. I think this is probably the issue? I'm not sure how AMBER handles the protein-gly links, but I'm guessing there are distinct CMAPs in this case?

[lohedges]

This should now be fixed in #401.

[BenTanYK]

Fantastic, thank you @lohedges !