GROMACS version:2018.8
GROMACS modification: Yes/No
Here post your question
I’m really new for gromacs. I have run a protein complex for 100 ns, and during my simulation, it jump over the box boundaries and flash to another side.
Here is the commend I run for box generation:
gmx_mpi editconf -f 1hyr_processed.gro -o 1hyr_newbox.gro -c -d 1.0 -bt cubic
I also fine some solution on the Internet, here is the commend I run for Trajectory processing:
gmx_mpi trjconv -f result/corrected.xtc -s md.tpr -o result/centered.xtc -center -pbc mol -ur compact
gmx_mpi trjconv -f result/corrected.xtc -s md.tpr -o result/fixed.xtc -fit rot+trans
They can run successfully , but they don’t work.
I don’t know if i depict the problem clearly.
The question is
- How can i fix this problem?
- How should i depict this problem universally?
Please feel free to ask any question about this topic, thanks
If your solute is a complex, you might need to use -pbc cluster to make sure the components of the complex are put in the same image. After that you can use -fit rot+trans for visualization, or do the alignment in your visualization software (e.g. Extensions > Analysis > RMSD Trajectory tool in VMD).
Whenever dealing with a complex, it is often easiest to center first on a single chain (create a custom index group for this) and then center again on the whole complex if you want.
Thank you so much for your help!
I have try to put complex in the same image by following commend:
gmx_mpi trjconv -f result/corrected.xtc -s md.tpr -o result/centered.xtc -pbc cluster -ur compact
gmx_mpi trjconv -f result/corrected.xtc -s md.tpr -o result/centered.xtc -pbc cluster -ur compact
I don’t know if these commend can put complex in the same image, but these commend didn’t solve the problem, can you give me some commend that can put the complex in same image?
By the way, when i need to choose a group, the following message will sent to the shell:
Will write xtc: Compressed trajectory (portable xdr format): xtc
Reading file md.tpr, VERSION 2018.8 (single precision)
Reading file md.tpr, VERSION 2018.8 (single precision)
Select group for least squares fit
Group 0 ( System) has 218876 elements
Group 1 ( Protein) has 8308 elements
Group 2 ( Protein-H) has 4263 elements
Group 3 ( C-alpha) has 527 elements
Group 4 ( Backbone) has 1581 elements
Group 5 ( MainChain) has 2111 elements
Group 6 ( MainChain+Cb) has 2608 elements
Group 7 ( MainChain+H) has 2623 elements
Group 8 ( SideChain) has 5685 elements
Group 9 ( SideChain-H) has 2152 elements
Group 10 ( Prot-Masses) has 8308 elements
Group 11 ( non-Protein) has 210568 elements
Group 12 ( Water) has 210561 elements
Group 13 ( SOL) has 210561 elements
Group 14 ( non-Water) has 8315 elements
Group 15 ( Ion) has 7 elements
Group 16 ( NA) has 7 elements
Group 17 ( Water_and_ions) has 210568 elements
is that mean my components of the complex are put in the same image.
Thanks again for your reply!
When saying “didn’t solve the problem”, can you explain what the outcome is? If you’re using -pbc cluster, you want to choose your solute (normally Protein) as the group for clustering. The message you sent (“Select group for least squares fit”) comes up when you choose the -fit option though so I’m not sure if that’s your problem.
I’m so glad you replied.
Firstly, I’m sorry I didn’t describe the problem clearly. In my simulation, the protein complex has three chains, and it will move towards the edge of the box such that one of the chains will cross the edge to reach and the other side of the box.
I run this commend
gmx_mpi trjconv -f result/corrected.xtc -s md.tpr -o result/centered.xtc -pbc cluster -ur compact
and choose Group 1(protein) for clustering and Group 0 (system) for output.
Then I run
gmx_mpi trjconv -f result/corrected.xtc -s md.tpr -o result/fixed.xtc -fit rot+trans
and choose Group 1 for fixing and Group 0 for output.(I’m sorry I paste the wrong commend in my earlier reply.)
I paste the group info. because I don’t know how to define the complex in the same image and “didn’t solve the problem” means the chain still move cross the box.
Thank you
Please try gmx trjconv -pbc mol -center with my strategy of centering on a single protein chain. It is generally a good solution. Clustering can sometimes break down in my experience.
Yes! I have try your strategy with the following commend:
gmx_mpi make_ndx -f md.gro
and choose ri 253-527 as Chain C
than I run
gmx_mpi trjconv -s md.tpr -f result/aaa.xtc -o result/centered.xtc -pbc mol -center -n index.ndx
and choose Group 18 (ChainC) for centering, Group 0(system) for output.
When I visualized the .xtc file via pymol, I was pleased that it looked like some changes had occurred.
But the box changed, too. It used to be a cube and now it look like this
Did I do something wrong?
What @jalemkul says is true, clustering can fail with elongated molecules where the reference atom/COM of one chain is far away from the other. In general, the less globular the geometry of a complex and the smaller the box relative to it, the more ways in which different commands might break down.
With your fixed PBC, it seems you start from a modified aaa.xtc, is the solvent in your starting trajectory distributed as a cube? It looks as if the box was rotated at some intermediate point, but it’s hard to guess.
try to make index file of the chain c .gro from energy minimisation.
Eg.
gmx make_ndx -f energy_minimisation.gro -o chain_c.ndx
yes, the starting trajectory is distributed as a cube
Thank you for your reply, I will try it tomorrow.
Thanks to each of you for your replies and Happy New Year 2024!
1 Like
