What you need is to convert your old.gro file into a PDB with

trjconv

Then you can make all the changes with the software of your preference to manipulate protein structures (for instance "rosetta suit", cut, etc. ).

Finally process the new.pdb with pdb2gmx

to recreate your

new.gro

new.itp

and

new.top

files

voilà!

Editing itp files is not simple and requires a good deal of experience, as every tine you insert or remove an atom you need to update all indexes in the parameter sections:

BONDS

PAIRS

ANGLES

DIHEDRALS

Usually, that requires writing programs to take care of the task.

It is not recommended to do such thing manually, unless you really know what you are doing.

Best wishes,

Rogelio

Hi Rogelio,

Thanks for your reply.

I can try using trjconv to go from .gro -> .pdb and then back again. Then hope using a third party software that the .pdb file isn't too different from the original .gro.

I was hoping to find something that would help edit the .itp file directly. If I can't find anything I'll probably have to write software that will do this for me.

There is no tool to edit itp files

If you need one, you would need to write the code it yourself.

The most complex modifications I have made to an ITP file is merging something at the end. Then you may add the corresponding relating those atoms to the rest of the molecule, with the corresponding changes.

I possible to have a workorround if you take advantange of AMBERTOOLS

and acpype

Prepare your system in tleap, where you can make changes with appropriate commands.

save the coordinates and the "prep" file.

give these two to

acpype.

acpype will create the gromacs itp and the gro file for you.

I still do not undertand why do you need to edit the itp files instead of going back to the *.gro to the PDB file and go forth again.

In answer to the question at the bottom of your post.

I am trying to resurrect an old project from a few years ago. The protein is massively modified with bespoke force constant data from QM calculations for metal binding sites and glycosylation bonds. Doing the .gro to .pdb approach, in theory will work, but will require a lot of playing around with residue entries to get pdb2gmx to be happy with it. Also, because a large number of the residues in the .gro file are new residues in the force field, third party software e.g., Maestro, Avaogadro, will not recognise the bond order and will probably mess up a modified .pdb file. Hence any attempt at making a .gro and .itp will blow up as pdb2gmx won'tbe able to go from .pdb to the FF files.

Consider all of this when all I want to do is replace one amino acid!

I see!

In that case, let me say that GROMACS does not care how your atoms are numbered or ordered a long as every bond, angle, pair and dihedral is parametrized correctly and declared. Than means it is possible to add a residue at the end and linked to something in the middle of the molecule, i.e. your file could have an altered residue which has added atoms at the end. For instance if you want to change a isoleucine to an aspartic acid, you will have to change some atoms in the isoleucine to dummy to disappear their charge and LJ interactions. Then add the missing atoms at the end. Their numbers will follow.

Declare the bond records linking the new atoms to the old residue.

Declare the angles. Declare new dihedrals and pairs.

You may use the aminoacids.rtp database to identify which new records are needed, or you may take a couple of small tripeptides and guide:

GLY-AA0-GLY

and

GLY-AA1-GLY

(where AA0 is the starting residue and AA1 the new residue)

and create their itp's using pdb2gmx.

It is hard work. It is somehow easy to make mistakes.

Finally, just in case it may help you to rebuild the metal centers:

http://ambermd.org/tutorials/advanced/tutorial20/mcpbpy.htm

Best wishes

Rogelio

Anthony Nash you should try with pmx module from Daniel Seeliger (https://github.com/dseeliger/pmx). pmx is a toolkit for free-energy calculation setup/analysis and biomolecular structure handling. With this module is perfectly possible to generate a dual topology (in this case alanine/glycine) for your system and run the proper simulation either with glicine or alanine on your protein. For instance, suppose alanine correspond to the state 1 and glicine to state 2. In the .mdp file you can control which of these aminoacids will appear in the protein by changing lambda parameter (i.e. lambda=0 --- alanine, lambda=1 --- glicine)...

only one question: you want to change alanine into glycine accross the MD simulation or simply you want run two different simulation (i.e. one with alanine and other with glicine)

hope this helps!

Hi Mario. Thanks for your reply. I've used PMX before to perform a full protein-length alanine scan of a molecule. It is very useful indeed. In hind sight, this would have been excellent. Yet, I am still hoping there is something out there that can perform the file changes (.gro & .itp) within the context of gromacs more as an ad hoc alteration to a system's structure than something I had planned for.

In reality I think will reluctantly battle through pdb2gmx.

I'm playing around with this further and I'll keep an update with how this has progressed. So far:

1) Removed the improper dihedrals from the aminoacids.rtp force field of the bespoke residues (they were failing at pdb2gmx as they were on the N/C terminal, to which unique N??? and C??? bespoke amino acids had not been calculation).

2) pdb2gmx whilst ignoring hydrogens (-ignh yes). This will make the .pdb to .gro pbd2gmx a huge pain, as getting the correct protonation state in the catalytic region is going to be very tricky to pull off by hand.

3) Made the amino acid mutations using Maestro.

Unfortunately, when you expect each modified structure into a .pdb, it's messed up relative to how pdb2gpmx will read each residue from my respective topology database.

In conclusion so far, Maestro is very easy to use, but the .pdb file saved is very different from the .pdb file loaded in and will require a lot of changes by hand (not even sure where to start), 2) and, something like a mutate_amino function in gromacs (not using the lambda functionality) would be amazing :-)

Hi Anthony,

As Rogelio suggested above, You may consider manual modifications in case you dont want to disturb the old files alot. Replace and/or add atoms, atomtypes, IDs and coordinates of Alanine (to be mutated) with required amino acid (Glycine) in the gro file, you may use external PDB coordinates (From Maestro) as reference to carry out the replacement. Further, you have to modify the ITP file manually for ATOMTYPES, BONDS, PAIRS, ANGLES, and DIHEDRALS. This can be time consuming but possible!

Hi Raj,

I know how to do it (I've done it now after hours of adjusting .pdb files manually), it was more a question of where, similar to AMBER, Gromacs had a function that I didn't know of which would change amino acids in a .gro and .itp format.

The closest is:

https://atb.uq.edu.au/

If you can build you "residue". As an isolated unit. You may find this site usefull.

It will provide a Itp file for some popular gromos forcefields.