Gromacs Output Help

[aab18011]

Hey,

I am working on a research project and wish to present my findings via molecular nodes, mostly because it looks amazing and allows me to really focus on things. Problem is, that I am working with gromacs 2024.2, but it seems to have too many molecules for MDAnalysis to work. I had previously saved my .gro file as a pdb file using amber numbering, but I found out that amber numbering is not supported.

Since then I have tried a few things, namely just the original .xtc and .gro files, which worked, but lacked the secondary structure. DSSP did not work either, as there were some nonstandard amino acids within. When I loaded this in ChimeraX or VMD, no issue with secondary structure.

If there is a nice way to convert or prepare the file (ie. separating the membrane, solvent, protein and cofactors into separate pdb files and trajectory files?), or any scripts I can write, please let me know.

For more information on the protein, size, membrane composition, etc.:

• 5HT2A receptor, from pdb 6A93, with all but Zinc 2+ cofactor removed
• Lipid bilayer, made with non-homogenous leaflets (POPC, POPE, DOPC, DOPE, Shingomyelin, Ceramide 3, Cholesterol) and (POPC, POPE, DOPS, POPS, PI(4,5)P2, Cholesterol)
• CYS modification to residue CYS397, a palmitoylation
• 7.4 pH with CHARMM-GUI autodetected protonation of HIS (HSP and HSD variants, about 4 total)
• Standard sulfide bridging
  -TIP3 water, with 0.15M CaCl2 (neutralizing)
  Total molecule count is greater than 120,000, which means an atom count over 800,000.

[BradyAJohnston]

Hey there! Should hopefully be able to help get this working.

The issues with GROMACS 2024 sound like something worth looking into, but if it is an MDAnalysis issue then it's worth taking that to the MDAnalysis repo with details there.

Molecular Nodes will let you have multiple trajectories loaded at the same time, so the easiest might be to split it into protein and non-protein atoms. That way you can use the .pdb file with secondary structure annotation for the protein component (open in pymol / chimera & save new file with secondary structure) and the .gro for everything else.

Secondary structure import for trajectories currently still isn't great so getting the information actually usable in Molecular Nodes currently is a bit tricky (apologies on that front).

The current solution is to import the .pdb file via the regular "Local" method (or via drag and drop) and ensure that the style is disabled and you don't remove solvent (if any nodes were added to the tree of this object remove them).

The in the node tree of the imported trajectory (for the protein), we can sample the sec_struct attribute from the other import method and apply it to the atoms in our updating trajectory like in the screenshot below. This won't update with the changes in the trajectory and are static from the topology file, updating secondary structure information is something that still needs to be worked on #872

[aab18011]

This was pretty much the answer. I have to re-arrange a few things for my own reasons, but this worked. I actually was able to use some 'gmx' commands to save just the protein in a pdb snapshot, as well as other portions of the protein (static/anchored regions that have low RMSF and low RMSD overall). The trajectory was still able to be used. Thanks a bunch!

Also, if you have a contributing section, I will take a look and see if I can help out with this.

[BradyAJohnston]

Glad to know that it's working!

There is a CONTRIBUTING.md but it doesn't quite cover everything as working inside of Blender means a whole bunch of weird and particular ways of working compared to normal python projects.

I'm always excited to get people on board & contributing though, so happy to answer questions / have discussions around new features that you would be interested in contributing.

[BradyAJohnston]

Additionally I would love to see results of what you are making. There is additionally a Discord Server that you can join and chat with other science / Blender viz people.