Large RMSD Differences of Homologous Proteins

I made the new .tpr file using gmx grompp, just as I did before running the final md simulation, but using the fixed.gro in the -c option. Was this correct?

Are there any other possible reasons other than the protein starting uncentered to explain the unusual RMSD?


Can you provide your input files somewhere, and the first few frames of your trajectory (in any format)? I don’t want to guess wildly about anything and I still have no idea how one would obtain such large values of RMSD immediately at the outset of the simulation.

Would there be a good way to transfer my files to you?

If they can’t simply be attached to your post here, upload to Google Drive and provide the link. Make sure sharing options allow for anyone with the link to see the contents.

In a Google Drive folder, I have uploaded all files used to calculate the rmsd, as well as the fixed.gro file made, the previous md_0_150.tpr, and an MPG of the filtered trajectory visualized with VMD.

The link to the Google Drive folder is here:

Please reduce the file size of the trajectory by only including the first few frames. I’m pretty sure I know what’s going on but I want to test something. 5-10 frames should be sufficient to diagnose.

I have uploaded a .xtc file of the first 10 frames to the same Google Drive I shared!

I can’t reproduce the weird numbers.

gmx rms -s fixed.gro -f md_0_150_chainA_frames_0_10.xtc


# This file was created Thu Oct 29 14:08:01 2020
# Created by:
#                       :-) GROMACS - gmx rms, 2020.3 (-:
# Executable:   /usr/local/gromacs/current/bin/gmx
# Data prefix:  /usr/local/gromacs/current
# Working dir:  /Users/jalemkul/Downloads
# Command line:
#   gmx rms -s fixed.gro -f md_0_150_chainA_frames_0_10.xtc
# gmx rms is part of G R O M A C S:
# Giant Rising Ordinary Mutants for A Clerical Setup
@    title "RMSD"
@    xaxis  label "Time (ps)"
@    yaxis  label "RMSD (nm)"
@TYPE xy
@ subtitle "Backbone after lsq fit to Backbone"
   0.0000000    0.0004864
  10.0000000    0.0865524
  20.0000000    0.0958802
  30.0000000    0.0962045
  40.0000000    0.0943499
  50.0000000    0.0967933
  60.0000000    0.0874779
  70.0000000    0.0879277
  80.0000000    0.0977267
  90.0000000    0.1061396
 100.0000000    0.1017297

These numbers are sensible and don’t look like what you have. Note that I’m selecting the protein backbone for the calculation. Did you do something different?

Thank you very much for trying to calculate it yourself, Dr. Lemkul. It is greatly appreciated!

I did select the protein backbone for the calculation. The only difference was that I used the centered .tpr file that is also in the Google Drive folder as opposed to the fixed.gro file that you used for the reference structure. I’m assuming that this would make the difference.

Did you manually account for the Mg ion mass in order to use the fixed.gro file?

I changed MG -> Mg in fixed.gro since atommass.dat works on a case-sensitive basis.

That appears to have done the trick when calculating for all the frames as well! The RMSD appears to hover around 0.25 nm.

Thank you so very much for your help!