Residue names changing after using minimizatioin

GROMACS version:2021.3
GROMACS modification: NO
Here post your question

I created up a membrane of DPPC.pdb in water using packmol and placed it in a box using editconf with output.gro. This looked as the pdb i.e. normal bileaflet in water.

Then ran a short minimization. The PE was high 10e7, but the system ran. What is very odd is that in viewing the DPPC-em.gro with vmd, the names and colors of the residues were switched. The water shapes were in the correct position but were named DPPC and were colored as carbon would be in vmd. The DPPC shapes were in the correct bileaflet position, but they were now colored red and white and had the name SPHE ( from atb ) as their residue name in vmd . Pymol also show the same inversion. The gro file of the system appeared correct - snippet below. Rebooting had no effect. Rebuilding in packmol had no effect.

I’ve never encountered this before and have no idea except perhaps the model is not correctly packed and minimization did what it could to do its job.

Any suggestions appreciated.

AMD Ryzen 7 2700
Linux MInt 19.2 Tina

20197SPHE H385989 5.941 15.579 2.402
20197SPHE O185990 5.830 15.565 2.327
20197SPHE H285991 5.759 15.650 2.339
20198SPHE H385992 5.783 15.468 2.339
20198SPHE O185993 5.859 15.585 2.164
20198SPHE H285994 5.771 15.567 2.108
20199SPHE H385995 5.908 15.681 2.156
20199SPHE O185996 5.970 15.473 2.117
20199SPHE H285997 6.059 15.428 2.218
20200SPHE H385998 5.999 15.371 2.293
20200SPHE O185999 6.110 15.515 2.269
20200SPHE H286000 6.135 15.361 2.169
20201DPPC H8086001 5.899 15.368 2.044
20201DPPC C4086002 5.840 15.406 1.960
20201DPPC H7886003 5.833 15.317 2.118
20201DPPC H7986004 5.975 15.297 2.007
20201DPPC C3986005 6.057 15.562 2.008
20201DPPC H7686006 5.986 15.601 1.935
20201DPPC H7786007 6.126 15.487 1.967

did you get any warnings concerning atom names/types when running gmx grompp to generate tpr file for the minimization?

No warnings, but encountering “fatal error there is no decomposition for x ranks compatible with a given box size…” This occurs when packmole packs too loosely.

In another raft model I create a 200 molecule (SPG ) as the raft and minimize/nvt/npt that with z restraint with no issues. the box might be 3 x 20 x20 nm. Then no matter how I pack, say DPPC, around this [ with packmole, with insert-molecules, or even with solvate ( a DPPC solvent box works nicely ) ], the system explodes on minimzation and the names are reversed in this case rdd is used.

The individual molecules 3 or 4 of each type run just fine so the itp files are correct

Happy to include any files if requested. I understand it’s difficult with out more detailed info.

Not to be too skeptical here, but are you sure that grompp did not issue any warnings? The only way species get renamed and give the behavior you are describing is if the topology and coordinates are in a different order. This can happen in complex systems. The extremely high potential energy also suggests some weird mapping between the two. Please provide the input (un-minimized) and output (minimized) .gro files for inspection. We can learn a lot from this.