[gmx-users] torsion angle scan

rob yang nextgame at hotmail.com
Mon Dec 12 16:41:36 CET 2005

i've always done option 2), the relaxed torsion scan, for both my qm and 
forcefield scans. intuitively i think the advantage is that your energies 
won't be dominated by unfavorable steric clashes this way. also, i think 
normalization of each of the qm and ff profile before subtracting them is 
important as well.
i'll look into ffscan to see if it allows me to do constrained minimization. 
thank you for that suggestion.


>From: Konrad Piwowarczyk <konradpi at poczta.fm>
>That is the topic that I am also interested in.
>It seemed for me that ffscan would be a good tool, at least after some 
>changes in its code, since the tool is designed rather for force-field 
>optimization. However I could not find *any* documentation about the input 
>format of its input (params.dat file particularly).
>So after all I prepare - just like You - a large set of structures and - 
>using bash script - run a single (1 step) md simulation for each structure. 
>That is not pretty but works.
>The other approach would utilise probably some kind of constraints but for 
>torsion scanning, positional constraints are not suitable.
>By the way, I am aware of 2 schemes of torsion fitting procedure:
>1. minimize a structure using QM at some torsion angle and make a set of 
>structures that differ only by the value of torsion. Just get single point 
>energy for each structure, you have total energy profile Eqm(theta). Than 
>one can compare it with the energy profile (for the same set of structures) 
>made by MD however with given torsional potential set to 0 -> Emd(theta). 
>The difference (with some constant bias) Eqm(theat)-Emd(theta) is just the 
>torsional profile that we need.
>2. Do the same but for each structure do additional optimization with given 
>torsion fixed. Is the resulting energy difference Eqm(theta)-Emd(theta) 
>still the torsion profile?
>Which one of this two should be used (the first one is of course much 
>simpler to do)?
>I would appreciate any comment on the topic from people that design new 
>forcefields using Gromacs - are there any?
>rob yang napisa³(a):
> > Dear list,
> > Is there a way to do systematic relaxed torsion scans in gromacs? In
> > particular, can I minimize a given structure with 1 bond frozen? I have
> > already generated a series of PDB files each differ by x degree around 1
> > torsion angle, so writing out pdb's is not a problem.
> >
> > thank you very much.
> >
> > rob
> >
> >
> > _______________________________________________
> > gmx-users mailing list
> > gmx-users at gromacs.org
> > http://www.gromacs.org/mailman/listinfo/gmx-users
> > Please don't post (un)subscribe requests to the list. Use the
> > www interface or send it to gmx-users-request at gromacs.org.
> >
> >
>gmx-users mailing list
>gmx-users at gromacs.org
>Please don't post (un)subscribe requests to the list. Use the
>www interface or send it to gmx-users-request at gromacs.org.

More information about the gromacs.org_gmx-users mailing list