I wanted to report something that occurred during my simulation, but was never mentioned (I never got any error message).
I have a protein-DNA complex and simulated this with the amberff14pbsc1 force field in Gromacs. After the runs everything seemed fine, but during the analysis I noticed that there is one base in my DNA which just went missing during EM/NVT/NPT runs and I was never warned during the whole runs.
Can someone tell me how something like this can happen and how I can check if my output is actually correct?
Have you tried checking the output .gro files and .top files to see whether the residues are still listed? I’ve found that features sometimes seem to disappear when looking through visualization software even if they are still accounted for in the actual simulation data.
I have tried to check, but I am not exactly sure how to find them in the files. Anyhow, the residue can not be found neither in VMD nor with pytraj, so I think it really is not there…
Could you send what the visualization looks like in VMD and which atoms were disappearing when you looked using pytraj? (I’m not familiar with using pytraj so am not sure what the output is)
Regarding how to check the files, if you open the topol.top file with a text editor (sublime text is what I’m using here), the DNA molecules would usually have atoms listed under different residues depending on the base they belong to (so in this case, atoms under residue 21 belong to an Adenine, residue 22 is a Thymine, etc.) For a .gro file, all the atoms would just be listed with the residue number and base listed in the first column to the left (residue 21 would then be 21DA, etc.) If you could find which atoms disappear at which stage of the simulation (the same information is also in the .pdb file if that’s what you started with), it might help point towards what the issue could be.
Here, I can not see the single residues and I do not have another one I guess (only the ones for each chain).
The Pytraj output just gives me one basepair less in the listing, this is why I started checking on it.
In VMD, it looks like the following:
You should be able to find the topol.top info for each DNA strand in topol_DNA_chain_C.itp and topol_DNA_chain_D.itp. Whenever the topology file is updated during these simulations, they always back up the previous copy by saving a version and putting a # to mark the previous versions. If atoms are disappearing from the simulation, it may be something to do with the .top file (which in this case is the .itp file for the DNA strand where the base pair went missing) or the .gro file.
Also, if you need a reference for which atom is which, use Mouse > Label > Atoms in VMD, and set ‘Drawing Method’ to CPK and you will be able to see which atom is which by clicking on them. The atoms in the DNA strand would be listed in sequence in the .itp file so that should help you identify which atoms disappear at which step of the simulation by comparing the files to the starting files.
Thank you for all your help, I will check again my topology files. It is just so wired that there is just one base which makes the error occur in pytraj is also not visible in VMD. Thats why I thought that iut must be a gromacs error…
But normally gromacs checks if the amount of atoms fits everything, doesn`t it?
Stuff can’t disappear. The image above means VMD is just having a hard time rendering a tube that fits a secondary structure for that geometry. It does that pretty often because it makes a lot of assumptions about what the structure should be. Your atoms aren’t disappearing anywhere. They can’t.
