Trying to run PENSA with AMBER files

[MauricioMartil]

Hi

I have been trying to run PENSA on my trajectories from AMBER. I generated the .psf file using parmed and used my own .incprd and .prmtop but whenever I try to run extract_coordinates.py, I am getting the following error message:

Traceback (most recent call last):
File "/media/mauricio/Expansion/POT1/pensa_pot_test/work/../../AF/a/pensa/scripts/extract_coordinates.py", line 29, in
extract_coordinates(args.ref_file, args.pdb_file, args.trj_file,
File "/media/mauricio/Expansion/POT1/AF/a/pensa/pensa/preprocessing/coordinates.py", line 44, in extract_coordinates
selection.write(out_name + '.prmtop')
File "/home/mauricio/.conda/envs/pensa/lib/python3.10/site-packages/MDAnalysis/core/groups.py", line 3756, in write
raise ValueError("No writer found for format: {}".format(filename))
ValueError: No writer found for format: traj/D51N_POT1.prmtop

I would like to know how I could make PENSA work using NetCDF coordinates.

[github-actions]

Congras on the first issue!

[martinvoegele]

Hi Mauricio, thanks for raising this issue!

It looks like you are trying to modify the function extract_coordinates() to write a PRMTOP file instead of GRO and PDB. Correct?

MDAnalysis — on which PENSA in-/output relies — only allows reading prmtop and psf files but not writing them. You can find a list of file formats supported by MDAnalysis here:
https://userguide.mdanalysis.org/stable/formats/index.html#coordinate-readers
Unfortunately, this limits the options to modify PENSA output.

Why would you like the output to be in NetCDF format? Depending on this, maybe there are other options to reach your goals.

[MauricioMartil]

Hi Martin!

Yes, I was trying to make some changes because I was not sure if PENSA only worked with XTC, GRO and PDB. So, from what I could see MDAnalysis should be fine with reading the NC files but I could see if I can use MDAnalysis or MDtraj to convert my PRMTOP and NC to GRO and XTC files.

I might need to add a line to the scripts so it can make that conversion. Sorry, I do not need the output to be in NetCDF, I think I did not expressed myself very well.

In the documentation, it is not very clear on how to make PENSA work with AMBER files, but also It could be that there is something different or wrong with my files.

On another note, I also got this other error:

Traceback (most recent call last):
File "/media/mauricio/Expansion/POT1/pensa_pot_test/work/../../AF/a/pensa/scripts/extract_coordinates.py", line 29, in
extract_coordinates(args.ref_file, args.pdb_file, args.trj_file,
File "/media/mauricio/Expansion/POT1/AF/a/pensa/pensa/preprocessing/coordinates.py", line 43, in extract_coordinates
selection.write(out_name + '.pdb')
File "/home/mauricio/.conda/envs/pensa/lib/python3.10/site-packages/MDAnalysis/core/groups.py", line 3674, in write
raise IndexError("Cannot write an AtomGroup with 0 atoms")
IndexError: Cannot write an AtomGroup with 0 atoms

Thank you for your time!

[martinvoegele]

Hi Mauricio, sorry for the delayed reply! I was traveling.

We have successfully used MDAnalysis with trajectories from Amber. But as far as I know, we always used topology files in PSF format, not prmtop or inpcrd.

Concerning the error message you report:
This error usually appears when your atom selection is empty. This means, most likely, that the selection string you provide to the function extract_coordinates() does not match any atoms in your system.
Since you use the script extract_coordinates.py I recommend you check the arguments --sel_base and --sel_file and the content of the file you supply to the latter. You can also manually modify this script to suit your needs.