Missing dihedral recognition using gmx chi

User discussions
1

GROMACS version: 5.1.2 and 2021.1
GROMACS modification: No

Hi all,

I am running gmx chi on various amino acids in GxG tripeptides to generate phi/psi and chi 1/chi 2 distributions of the x amino acid. Running into an issue with Valine. Here is my command:

gmx chi -f pep_ion_md.xtc -s pep_ion_md.tpr -o chi_output_50_300/order.xvg -rama yes -b 50000 -e 300000 -dt 2

In the output, it recognizes the structure has 9 dihedrals (3 pairs of phi/psi dihedrals and 1 chi dihedral for valine) but then for some reason does not recognize the chi dihedral for valine:

3 residues with dihedrals found
9 dihedrals found
Last frame 125000 time 300000.000
j after resetting (nr. active dihedrals) = 8
WARNING: not all dihedrals found in topology (only 8 out of 9)!

No chi1 & chi2 angle for GLY1

No chi1 & chi2 angle for VAL2

No chi1 & chi2 angle for GLY3

This has not occurred for similar amino acids (Leu, Ile). I tried in two separate versions as indicated at the top of the issue. The simulation itself was prepped and run with Gromacs 5.1.2. I have simulations of GVG in multiple force fields and this is a force field independent issue. Gromacs version aside, it’s not clear why the topology would not have the extra dihedral term.

Any help is appreciated, or alternative way to gather the chi dihedral data for the amino acid from my trajectory.

Thanks,

Brian

2

I generated a topology file with both v5.1.2 and v2021.1. Both topology files have the dihedral listed which include atoms 10 12 14 16. I’ve included the gro file and the topology files generated using both 5.1.2 and 2021.1. They both include the first chi dihedral on line 271 in these files. So, I’m not really sure what this warning is about:

j after resetting (nr. active dihedrals) = 8
WARNING: not all dihedrals found in topology (only 8 out of 9)!

See files below:
pep.gro.dat (1.6 KB)
pep-v5-1-2.top (10.5 KB)
pep-v2021.top (10.1 KB)

3

Last bit of detail, I get the same warning when running gmx chi on Leucine:

3 residues with dihedrals found
10 dihedrals found
Reading frame 124000 time 298230.000
j after resetting (nr. active dihedrals) = 8
WARNING: not all dihedrals found in topology (only 8 out of 10)!
No chi1 & chi2 angle for GLY1
No chi1 & chi2 angle for GLY3

However it says it found chi1 and chi2 for LEU2 after the warning. So it seems that this command does not work for amino acids with only 1 chi dihedral?

Either way, I ended up using the mdtraj python package to complete this analysis for Valine (MDTraj — MDTraj 1.9.8.dev0 documentation).