gmx_MMPBSA failing when trying to run on a cubic simulation

[Kirtley Amos]

Hello all!

I hope that this message finds you well!

I'll start by saying that the gmx_MMPBSA package is great for our use cases. It's done exactly what we would like it to. However, I've encountered a problem with one of our simulations that causes gmx_MMPBSA to fail.

This is the error that I get:

return atom, words[3], int(words[2])

ValueError: invalid literal for int() with base 10: '36A'

This sim was built in a cubic shape and was fit with the following commands:

gmx trjconv -s topol.tpr -f xxx.xtc -o xxxx_center.xtc -center -pbc mol -ur compact

gmx trjconv -s topol.tpr -f xxx_center.xtc -o xxx_fit.xtc -fit rot+trans

This has worked for charmm-gui outputs and a built from scratch dodecahedron system so my question is this. Have I fit the cubic system incorrectly and that's the reason for the error or is it something else that I have missed? Or, is it that gmx_MMPBSA doesn't support cubic sims?

Any information would be useful as I would still like to glean information from this simulation rather than not use it.

I'm looking forward to hearing back from you all.

Best,

Kirtley

[marioe911116]

We are glad that you find gmx_MMPBSA useful... :)

seems like you have insertion codes in your system... could you please update ParmEd and see if that solves the error?

for conda environment + pip installation:

python -m pip install git+https://github.com/ParmEd/ParmEd.git@16fb236

or if you have AmberTools compilation + pip installation:

amber.python -m pip install git+https://github.com/ParmEd/ParmEd.git@16fb236

hope this helps!

[Kirtley Amos]

Hello Mario,

Thanks for getting back to me so quickly.

I'll have a look at this and see if it fixes the problem.

All the best,

Kirtley

--
You received this message because you are subscribed to a topic in the Google Groups "gmx_MMPBSA" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/gmx_mmpbsa/b7vHZRISMMY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to gmx_mmpbsa+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gmx_mmpbsa/37657247-ffcc-423b-bf4a-77bd0b9f0a8dn%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Kirtley Amos]

Hello Mario,

So I've tried the method that you mentioned to correct the error but I'm running into issues updating ParmEd.

Installing collected packages: ParmEd
  Attempting uninstall: ParmEd
    Found existing installation: ParmEd 3.4.1
ERROR: Cannot uninstall 'ParmEd'. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a partial uninstall.

I've gotten this error multiple times while trying to install.

Any thoughts?

Best,

Kirtley

[Mario Sergio Valdes]

When AmberTools is compiled, two versions of ParmEd are installed, one in /path/to/AmberTools/lib/parmed and one in /path/to/AmberTools/miniconda/lib/python3.x/site-packages/parmed. In this case, since the PYTHONPATH variable is exported, the first one is used. Simply go to the /path/to/AmberTools/lib/parmed folder and manually delete it and the associated *.egg file. Then reinstall parmed as above

HTH!

Mario S.

[Mario Ernesto]

Which installation did you have!? Conda environment + pip or ambertools compilation + pip!?

To view this discussion on the web visit https://groups.google.com/d/msgid/gmx_mmpbsa/17d0022a-a59b-4f5d-a184-b110a7181462n%40googlegroups.com.

[Kirtley Amos]

Okay,

Great. I'll do that. I installed within A conda Environment + pip. So, should i still go to the first path you listed, the shortest of the two without python and delte the *.egg file?

Best,

Kirtley

[Mario Sergio Valdes]

If you installed in a conda environment this should not happen, since the packages are installed using pip or conda. This problem usually happens when installing a package with python setup.py install, such as when compiling AmberTools. To be sure, show me the output from these command lines

echo $PYTHONPATH

python

import parmed

parmed.__file__

[Kirtley Amos]

Okay, so we found a workaround to update ParmEd and when we do so I still get the same error on the system of interest. It was upgraded from 3.4.1 to 3.4.3.

return atom, words[3], int(words[2])
ValueError: invalid literal for int() with base 10: '36A'

Error occurred on rank 0.

So, in order to upgrade we had to exclude the source in the bashrc referencing amber20. But, as you can see. We get the same error as before.

Thoughts?

Kirtley

[marioe911116]

Hello there! at this point is better to send around the files you are using to discard the possibility of the bug in the code... please, send the structure files, trajectory (only a few frames) and the topology files (if you are defining the topology)

cheers!