Trajectory slicing changes 3D coordinates

User discussions
,
1

GROMACS version: 2021.1
GROMACS modification: No

Due to large size for visualization, I sliced the trajectory during the last round of gmx trjconv using -pbc cluster. When the trajectory sliced into 0-250ns and 250-500ns, no problem occurred. However, when sliced into 200-500ns, it resulted in 16537 abnormal 3D coordinates, such as:

Blockquote
-1.795-1104908288.000 0.883

Overall, I took the following steps to process the trajectory:

  1. gmx trjconv -s sys.tpr -f sys.xtc -n index.ndx -pbc nojump -center -o nojump.xtc
  2. gmx trjconv -s sys.tpr -f nojump.xtc -n index.ndx -pbc cluster -center -ur compact -o cluster_200-500ns.gro -b 2e+05 -e 5e+05

Then after loading the output into VMD:
pbc wrap -all -centersel “protein” -center com -compound residue

The final output was saved from VMD after selecting the protein and water within 2.5 from the protein.

2

I would suggest analysing this in steps, to see where the problem occurs. I presume that the coordinates are OK after step 1, right?
Are they correct after step 2? Otherwise, try to divide this into multiple steps, just to be on the safe side, e.g.,

gmx trjconv -s sys.tpr -f nojump.xtc -n index.ndx -pbc cluster -center -ur compact -o cluster.xtc
gmx trjconv -s sys.tpr -f cluster.xtc -n index.ndx -o cluster_200-500ns.gro -b 2e+05 -e 5e+05

If there is a problem after this, try

gmx trjconv -s sys.tpr -f cluster.xtc -n index.ndx -o cluster_200-500ns.xtc -b 2e+05 -e 5e+05

to see if it works if writing xtc instead. Or would there be a reason why you need .gro output? If the coordinates are fine at this point, the problem is in the VMD transformations. I wouldn’t have any good suggestion how to address that, but at least the problem would be isolated.