[gmx-users] afm pulling

[Emily Walton]

Hi,

> Dear all,
> I am simulating an unbinding process with the pull code (afm option).
> I've been reading previous posts about this option and I just  
> wanted to
> clarify some ideas and ask some questions. There are:
> 1-The pulling can be done choosing two different molecules. Why a
> particular group within a molecule can not be used? Not even as a
> reference group?

Any group that is defined in your index file (specified with mdrun - 
pn pull.ndx) can be used. Look at made_ndx in the manual to learn how  
to add new groups.

> 2-Should I need to fixed the center of mass of the protein to avoid
> shifts? How can I do that? Would it be OK to use posre for the  
> backbone
> of the protein with carrying out the pulling?

It depends on what you want to do. Typically, something would need to  
be fixed. In ligand-protein systems, the center of mass of the  
protein is a reasonable choice. It was used by Grubmuller et al. in  
their 1996 Science paper on the unbinding of biotin-streptavidin.  
(Science, 271:954–5, 1996) For unraveling of proteins, one terminal  
atom may be fixed while the other is pulled.

Fixing the center of mass of a group in the index file can be done in  
the .mdp file you feed to grompp. See the manual, chapter 7.

> 3-Should I removed all the constraints for the bonds (SHAKE, lincs)?

I think this is a matter of preference. Some people do, some people  
don't. Again, it will depend on your system- is it stable without the  
constraints?

The bonds that are broken with afm pulling are bonds described by non- 
bonded interaction terms in gromacs. These bonds wouldn't be  
constrained by shake or lincs anyway, so the presence or absence of  
constrains should have a minimal effect on the simulation results.

> 4-once I have the pull.pdo the calculation of the applied force  
> will be
> F= K*(zspring -zgroup) (assuming I am pulling in the z direction)?

Yes. The spring is an ideal spring, so force is k*(extension). The  
extension can be calculated from the pull.pdo file.


-Emily Walton