LINCS error for isopeptide bonds

User discussions
, ,
1

GROMACS version: 2023
GROMACS modification: No

Dear Gromacs user community,

I have the following problem in Gromacs.

I am studying a protein that has an isopeptide bond. I went through the previous post on this forum in order to create the isopeptide bond. The method I followed is as follows.

  1. First, I placed the two proteins in a way that they do not overlap, and the residue of each protein (LYS and GLY), which will make the isopeptide bond is within the cutoff.
  2. I defined the corresponding LYS and GLY residue in the rtp, residuetypes, hdb, r2b files, which will make the isopeptide bond. I have successfully created the bond. I employed the amber a99SB-disp for building the topology. Solvated the structure in water and added NaCl to have a molarity of 150mM.
  3. Minimized the structure using SD and CG procedures.
  4. Run NVT for 1 ns with restraint (1000KJ.mol-1) on the protein C-α atoms. V-rescale temperature coupling was used with an integration time step of 2fs.
  5. Run 5 NPT simulations for 1 ns each, slowly reducing the restraint on the protein (1000,500,100,50,10 KJ.mol-1). Parrinello-Rahman pressure coupling was used.
  6. Run production simulation in NPT without any restraint.

Now, when I start the production simulation, after a few ns, LINCS error occurs on the isopeptide bond. I looked into the previous Gromacs post on LINCS error and tried to implement what has been suggested before. Clearly, the starting structure is not an issue here. I used Gromacs 2023 latest version with GPU implementation. The surprising thing is that when I run my system using only the CPUs, the LINCS error does not appear. I thought I had built the software wrongly. Then I installed Gromacs 2022.3, and the same problem appeared in the production simulation for the GPU version.
What could be the possible resolution? Is this a bug, or have I done something wrong?

Any help is appreciated.

Thanks,
Supriyo Naskar

2

I forgot to mention that when I am making the isopeptide bond, it is not creating any improper dihedral for the specified bond.
So, I manually added those impropers.
However, I am getting the LINCS error for both with improper or without improper dihedrals after a few tens of ns.

What could be the reason?