[Question]: [ERROR ] MMPBSA gmx trjconv failed when querying ../md_clean.xtc.

[zxianren]

My Question is...

Dear developer, I have run a MD which comprised three chains A, B, D. And i succeesfully caculate the AD pbsa, but when i run the bd, it seems to show gmx trjconv failed when querying ../md_clean.xtc. But i just used xtc file for ad pbsa.

[zxianren]

successful_AD_gmx_MMPBSA.log

[zxianren]

error_BD_gmx_MMPBSA.log

[zxianren]

I notioned the ohther users had meet the same error, i curious about if the gromacs gmx and the mmpbsa, what should i fix? Hoping your kindly advice.

[Valdes-Tresanco-MS]

Please, attach your files to check them.

[zxianren]

files.zip

Please, attach your files to check them.

Dear developer, here are the files, thanks for your kindly reply

[zxianren]

Please, attach your files to check them.

here is the link of xtc files.
https://drive.google.com/file/d/1Iu04Ol2LfEBVIHDi2lt4noLjPKexOEnX/view?usp=drive_link

[Valdes-Tresanco-MS]

In this case, the problem is that bd.tpr wasn't used to generate this trajectory. According to the atom count, bd.tpr has

Command line:
  gmx make_ndx -f bd.tpr


Reading structure file
Reading file bd.tpr, VERSION 2023.3 (single precision)
Reading file bd.tpr, VERSION 2023.3 (single precision)
Going to read 0 old index file(s)
Analysing residue names:
There are:   547    Protein residues
Analysing Protein...

  0 System              :  5450 atoms
  1 Protein             :  5450 atoms

while the trajectory has.

Command line:
  gmx check -f md_clean.xtc

Checking file md_clean.xtc
Reading frame       0 time    0.000   
# Atoms  121974

This causes Gromacs to terminate abruptly with a Segmentation fault, which is a non-descriptive error, so nothing is shown in the log. You can verify this by running the commands shown in the log file in your terminal.

Here are two options:

1. (Recommended) Use the same path for both cases, creating the groups according to the chains, where you can calculate A+B<>D, A<>D, B<>D (A<>B doesn't make much sense if they are the chains of an Ab, but it can be done without problem).
2. Create trajectories from scratch using the corresponding structures.

Keep in mind:

1. For Ab, separating its chains is pointless because simulating the independent chains can yield unrealistic results.
2. Never create tpr or any other derivative to compute on the same trajectory you already have; this leads to serious problems that are difficult to trace.
3. gmx_MMPBSA, since it uses a gromacs index, is quite flexible, but keep in mind that having rare, incomplete, or PBC structures can lead to errors in energy calculations.
4. It's always recommended to have well-defined chains in protein-only simulations. This helps to separate the groups correctly, distinguish them as entities in the topology and simulation, and visualize them more easily.

[zxianren]

Dear Dr. Valdés,
sorry for reply late. Thanks for your kindly reply. I rerun the MD on the weekend. And I follow your suggestions to generate the correct AD tpr, xtc and index file, and run the gmx_mmpbsa AD chains correctly without error. But when I use same input xtc, tpr and pdb files to generate BD, it happens the same error, couldn't gmx trjconv the bd,xtc? Here is the file linker:

https://drive.google.com/drive/folders/1rlKBns_1xjwCrUVjwN7mJExTR1RxSXTu?usp=drive_link

I aslo try to generate bd.xtc by using the pdb or tpr file for the corresponding structures, all hapended the error:failed when querying bd.xtc. And I check the atoms is 5450 in xtc and tpr files.
could you check the error again if you're available time. Thank your very much for your kindly and detailed reply. Best regards!

[Valdes-Tresanco-MS]

As mentioned above, it's not necessary to create files to process different parts of the same system. Based on the files you've shared, you're creating an in-place topology to calculate the interactions between A<>D and B<>D. This works for the first because the indices of the residues in the index_mmpbsa.ndx file aren't changed, but when processing B<>D, gromacs assumes the indices are maintained, which isn't the case, because in the tpr, the chain A is no longer indexed.

Therefore, to compute the interactions between A<>D and B<>D, you must do the following:

1. Ensure that the groups in index.ndx are correctly defined. According to my experiments, chainA (ri 0-197 -> 3141 atoms), chainD (ri 198-366 -> 2645 atoms), and chainB, the difference between group 1 and the union of the previous ones (in my case, 12|13 [chainA_chainD ->14], 1&!14 -> 2805 atoms). This is my index.ndx

 10 chD                 :  2645 atoms
 11 chA_chB             :  5946 atoms
 12 chain_A             :  3141 atoms
 13 chain_D             :  2645 atoms
 14 chain_B             :  2805 atoms

2. Select interest groups. If you want to compute A<>D, then the command should be:

mpirun -np 8 gmx_MMPBSA -O -i mmpbsa.in -cs md_prod.tpr -ct md_clean.xtc -ci index.ndx -cg 12 13 -cp topol.top -nogui

If it is B<>D, then:

mpirun -np 8 gmx_MMPBSA -O -i mmpbsa.in -cs md_prod.tpr -ct md_clean.xtc -ci index.ndx -cg 14 13 -cp topol.top -nogui

Note that you must use the topology, tpr, and production trajectory that contain all the atoms in the system. It is not necessary to create derivatives or modify any existing components to compute these interactions.

[zxianren]

Dear Valdes,
sorry for reply late. I follow your kindly advice to generate ndx file by pdb or tpr. And No redundant operations in tpr, xtc files. but command: mpirun -np 1 gmx_MMPBSA -O -i mmpbsa.in -cs md_prod.tpr -ct md_clean.xtc -ci index.ndx -cg 11 10 -cp topol.top -nogui #11 is chain D 10 is chain A, it error same as before:The complex index has fewer atoms than the topology. Same error happens as calculate chain DB, When i delete the chain B in topol file to make a new topolAD file, it calculate correctly like before: mpirun -np 1 gmx_MMPBSA -O -i mmpbsa.in -cs md_prod.tpr -ct md_clean.xtc -ci index.ndx -cg 11 10 -cp topolAD.top -nogui . But when I calculate the BD PBSA by using topolBD file, it also error. Here is the test file linker:
https://drive.google.com/drive/folders/1l9eOVRoIPtOa757tikYJsD3MUHMmJxKW?usp=drive_link
Thank you for taking the time to help me resolve this issue despite your busy schedule. Best regards！