[gmx-users] Can the structure built by Amorphous Cell of Materials Studio be used for GROMACS inputs?

chris.neale at utoronto.ca chris.neale at utoronto.ca
Sat Aug 25 20:44:39 CEST 2007


> Hi, everybody,
> I am a freshman for using GROMACS. I used to do simulations using  
> Materials Studio. Now I notice that versatile analysis tools are  
> present in GROMACS, so I decide to turn for GROMACS. I have built  
> some polymeric structures using the Amorphous Cell module in  
> Materials Studio4.0. Can these structures be directly input into  
> GROMACS for a further simulations? How can I do that? Can you give  
> some hints?

If you only want to use gromacs for analysis, then this can be done  
relatively easily. If, however, you want to use gromacs to continue  
simulations then it will be quite a bit harder.

I have never actually tried the procedure that I outline below, but I  
am fairly sure that it will work. At least I am convinced that, if it  
works, the analysis will be correct with the exception of analysis  
programs that do mass weighting, electron density calculation, or  
other charge-based calculations. If you want those things to work  
properly too then you will need to be more careful about the atom type  
and/or charge that you apply in the .rtp file (see below).

For the case of analysis first:

Output your MS4 trajectory as a single pdb file with all of the frames  
in one file (you'll probably need to use MS4 to do that (call this  
file traj.pdb). This can then be input to gromacs analysis tools. Also  
create a single frame pdb from the first frame (call this file  
first.pdb).
You will need a topology file in order to carry out analysis. However,  
for this purpose, the parameters don't need to be correct or even  
exist. All you need is the bonding profile and atom names. For the  
purpose of an example I have created an entry for molecular oxygen to  
a .rtp file:

[ oxygen ]
   [ atoms ]
   ;name  type       charge  charge_group   ;;this is a comment line
     O1   opls_154    0.0    1
     O2   opls_154    0.0    1
   [ bonds ]
     O1 O2

Take a look at some of the amino acid examples to see how residues are  
bonded to each other so that you can define monomeric building blocks.

Don't use this for simulations, but it should allow you to carry out  
analysis on trajectories that you already have. I chose opls_154 as an  
existing oxygen atom type, but as long as the type exists I don't  
imagine that it even needs top match the type of atom (i.e. you could  
use opls_154 as the atom type for every atom in your entire molecule).  
Assign all charges as zero, and put them all in a single charge group.

For a more complicated / realistic example take a look at the ethanol  
entry in the gromacs/top/ffoplsaa.rtp file, or the entries for amino  
acids in that same file.

Once you have correctly created your .rtp entry and placed it at the  
bottom of gromacs/top/ffoplsaa.rtp, then run these commands:

pdb2gmx -f first.pdb -ff oplsaa -p topol.top
touch empty.mdp
grompp -f empty.mdp -c first.pdb -p topol.top -o analysis.tpr

You will need to use the -maxwarn input with a large number probably  
because a lot of interactions will not end up being specified.

You now have a tpr file that can be usef for analyzing your traj.pdb file.

##########

If you want to actually continue your run then you need to do the same  
thing, but the charges and atom types need to be correct. Also the  
ffoplsaanb.itp and ffoplsaabon.itp files need to contain parameters  
for all types of interactions for your specified atom types. Further,  
you will need to specify your improper dihedrals explicitly in the  
.rtp file.

Note that for actual run continuation this is a huge amount of work. I  
don't reccommend it for a first time gromacs user.

##########

The actual parameters are not going to matter for analysis of  
trajectories that you already have. However, for continuation of  
simulations, please note that it is possible to make up a topology  
that runs without error and does not crash and yet gives you totally  
bogus results.

I realize that this post is a little scattered. If you do decide to go  
this route, please use the search feature of the mailing list (don't  
forget to check the mailing list box below the search button) and the  
manual to help you figure out what .rtp files are etc.

Chris.





More information about the gromacs.org_gmx-users mailing list