How to use output from PRODRG

GROMACS version: 2018.3
GROMACS modification: No

I am studying protein-ligand interactions, and my question has two parts which may be related:

  1. To get an .itp file for my ligand, I submitted its .pdb file to PRODRG server (since I am using Gromos forcefield), and I guess I should use DRGGMX.pdb and DRGGMX.itp.

How do I use the output files from PRODRG for Gromacs simulations?
I did editconfig to convert the DRGGMX.pdb into .gro, and then copied the content into the protein.gro
Then I added some lines to .top file about the ligand. Is this the right approach?

  1. How should I incorporate AutoDock result in the process?
    I tried running AutoDock to get my ligand’s best binding location before and after using PRODRG server. In both cases, creating a simulation box resulted in complete dissociation or weird structure of the ligand.

Don’t use PRODRG. Its topologies are demonstrably bad. https://pubs.acs.org/doi/abs/10.1021/ci100335w

Use ATB instead.

Preserving the ligand coordinates from the docked pose is sufficient for starting the simulation. Follow the general procedure at http://www.mdtutorials.com/gmx/complex/index.html

If the ligand is dissociating from proper starting coordinates, that means the ligand topology was bad. If you’re using PRODRG, that’s probably the reason (see the paper linked above).

1 Like

Thank you.

I tried using ATB to create .itp file for my AutoDock-ed ligand.
Then I tried combining it with my protein into one .gro file.

I found the position of the ligand to be different from that was calculated by AutoDock.
Does that mean I should create .itp file for the ligand before I run AutoDock simulation?

And, does an .itp file contain coordinate information? I guess it doesn’t?

Been a while since I used ATB but it’s probably performing a QM geometry optimization before computing parameters. There are no coordinates in a topology file, but you can use the .itp file it gives you with your existing coordinates, rather than the ones (apparently) changed by ATB.

Would it be easier to use the coordinate file given by these servers? Sometimes they change the atom names and by using their files you wouldnt have to manually edit atom names in your original coordinate file.

ATB server seems to change the order and naming of atoms to some extent, so yes, I think it’s better to find the binding site first (if you are doing docking simulation before Gromacs), submit the .pdb file to the server, and use their files.

ATB outputs your original .pdb file rearranged using the server’s format as well as geometry-optimized .pdb file (which completely changes the coordinate, so I don’t recommend using it).

So if the server provides you with its version of your coordinate file, then I agree that it’s better to use it, instead.

Cool! That makes sense… ive been using the output coordinate files recently because i didnt need exact binding sites.

Is the ATB server dead now …?

Works fine for me right now.

Oh! Thanks …
I guess the link given in your tutorial (Protein-Ligand Complex) for ATB doesn’t exist any more …
this one works
https://atb.uq.edu.au/

Updated, thank you.