GROMACS version: 2023.3
GROMACS modification: Yes/No
Here post your question
Hello everyone,
I’m new to GROMACS and I’m trying to understand how to use the output from my first simulation as the input for a new simulation with a different ligand. Specifically, I have a couple of questions:
Should I create a new topology file for the new ligand, or can I modify the existing one?
When preparing for the new simulation, should I use the “md.gro” file from my previous simulation, or would it be better to use a “start.pdb” file obtained from “trjconv”?
Any guidance or advice on these points would be greatly appreciated!
In general, you will need a new topology and build a new box. The output from the previous simulation has the wrong ligand, and the new ligand will have a different topology as it is a different molecule. You can use the output from the previous sims as a template or guide if, for example, you want to put the new ligand aligned with the old one, or if the new ligand is very similar to the old one and you are able to modify the topology by hand (e.g., just a protonation state). But in general the new simulation box and new topology will be different by definition, as you are changing a molecule inside the box.
Thanks for your reply. Let me explain about what I am looking for. I am working with two ligands: one is a sugar, and the other is a pigment. Although these ligands are structurally different, they are related in the sense that the first ligand might alter the system, induce a specific protein conformation, or influence the protein’s behavior, which in turn could impact how the second ligand interacts with it. After extensive research, I realized that I am unable to directly create a combined topology for the “protein + first ligand.” As you mentioned, I might be able to use the output from the first simulation as a template or reference.
However, I have several questions regarding the process. Should I extract the structure using trjconv—and if so, should I use the first frame, last frame, or an average structure? After that, do I need to separate the protein, generate a new topology, and then combine it with the topology files for the original first ligand and the new second ligand? How should the positions of the first and second ligands be determined? My docking results suggest different positions for the second ligand depending on whether I use the original protein structure or a structure extracted from the simulation (for example, the structure at time step 0).
I am not an expert in ligand binding, so I hope someone else can chip in and share, and in any case I advise you to refer to similar works in literature.
My two cents are that you do not need a new topology if the protein changes conformation. As long as the protein is the same, whether or not some kind of binding pocket opens or closes etc, the topology of the protein won’t change. Same thing for the ligands. If the ligands binds an opens some kind of other pocket - here you are the expert of your system - then at the same time it doesn’t make sense to generate a topology of a protein + ligand, as they are still two separate molecules. You could take your protein + ligand 1 system and add ligand 2 in its binding site. Or you could take just the protein and restraint some of the atoms to keep it in the conformation induced by ligand 1. My main question however would be: do you have any experimental data that show the binding pocket and the binding pose? Or do you have only docking and no knowledge of the pocket? And what do you want to simulate? A binding event/estimate the binding free energy?