I am learning how to use GROMACS and I am interested in protein-protein interactions.
I found the GROMACS tutorials very useful and I successfully went through them, however, I tried to adapt them for a protein dimer I am working on, and I found it somehow more challenging.
Can you suggest any tutorial that explains step by step how to run GROMACS with two proteins?
Is it also possible to use AlphaFold structures to do so?
Thanks for helping
Paolo
GROMACS version: GROMACS/2023.3-foss-2022a-CUDA-11.7.0
GROMACS modification: No
Can you first explain what you found challenging? Any errors you encountered? In principle having two proteins isn’t much different from having one, unless there’s something very peculiar about the desired setup.
Gromacs won’t care if your structure comes from a crystal, cryo-EM or AlphaFold, that’s a question of validity of the starting structure, and it’s hard to give a once-and-for-all answer without any details.
Create a two chain for the protein you wanted to simulate and pass through gmx pdb2gmx
You will have the topoogy and system for the two proteins with different chain ID. Remaining procedures are same as other system
Which one(s) specifically? With which force fields? Different FFs might use slightly different naming conventions, and not always capture all the variability. It can also be that you’re not distinguishing chains in the input.
Protein Q9ZNV6
Program: gmx pdb2gmx, version 2021.3-bioconda
Source file: src/gromacs/gmxpreprocess/pdb2gmx.cpp (line 790)
Fatal error:
Atom HD1 in residue HIS 169 was not found in rtp entry HIE with 18 atoms
while sorting atoms.
I have found HIS in residuetypes.dat,but none atoms in aminoacids.rtp
HIE is ε-protonated histidine, but the presence of HD1 means δ-protonated histidine. The easiest solution is to use -ignh and have pdb2gmx rebuild appropriate H atoms, otherwise you need to revisit why you are calling a residue HIE when the coordinate file says it should be HID based on protonation. If that is experimentally assigned, you will want to dig into that more.
No, neither of those options are appropriate as they serve different functions. There is almost never a valid case to use -missing because it will omit atoms that aren’t H, leaving incomplete topologies.
That sounds like an entirely separate problem, with atoms named incorrectly across the entire structure, or missing entirely. Please provide the full terminal output from pdb2gmx, including your command.
Hello all, when I conduct protein-protein interaction after finishing gromac tutorial of protein-ligand interaction, I stuck at here. In the tutorial, it writes:The ligand introduces new dihedral parameters, which were written to “jz4.prm” by the cgenff_charmm2gmx.py script. At the TOP of topol.top, insert an #include statement to add these parameters:
; Include forcefield parameters #include “./charmm36-jul2022.ff/forcefield.itp”
; Include ligand parameters #include “jz4.prm”
[ moleculetype ]
; Name nrexcl
Protein_chain_A 3
However, when the other ligand is another protein but not samll molecules, how could I add this parameter(since there is no proteinB.prm after I conduct both protein using the following command: “gmx pdb2gmx -f proteinA_clean.pdb -o proteinA_processed.gro -p protein.top -ter” and" gmx pdb2gmx -f proteinB_clean.pdb -o proteinB_processed.gro -p proteinB.top -ter"
It seems that the differences of the force fields. AMBER99SB-ILDN can’t work well , but CHARMM36 seems work well.
However, there is still the problem of atoms in the subsequent charmm36:
Have you tried compiling the latest GROMACS version from source? I have often encountered weird bugs from poorly compiled packages. Worth a try. These kinds of simple errors should be easily solved with -ignh and if they aren’t, something more fundamental is going wrong.
