Hi Jessica,

Playing with the vdw force constant and size scaling is almost always  
a bad idea.  You simply end up with low-energy structures that badly  
fail a procheck style verification.

If you're having trouble converging, you might consider starting with  
fewer degrees of freedom.  Fix your backbone torsion angles to their  
values in your model, anneal, and then allow them to move and re-anneal.

Force constant scaling is generally used only with energy terms that  
can take on very high values when you're a long way from the correct  
structure. The NOE term is the classic example of this. The games we  
play with the vdw term exist to increase conformational sampling,  
since the vdw term is ordinarily the primary hindrance to effective  
sampling.

Hope this helps.

--JK

On May 12, 2009, at 4:38 PM, jlgiffor at ucalgary.ca wrote:

> Hi,
>
> I am attempting to calculate a complex structure using a very similar
> starting model, RDCs, intermolecular NOEs and two-step low temperature
> simulated annealing. Can I adjust the force constants in the script,
> particularly those for van der waals repulsions and dihedral  
> restraints to
> decrease the number of van der waals and dihedral angle violations.   
> If
> so, to what values or types of values should I adjust them to? What  
> are
> things I should consider when I think about this?  Also what is the
> advantage of ramping a force constant instead of leaving it static  
> at the
> final value?
>
> Thank you
>
> Jessica Gifford
>
>
> _______________________________________________
> Xplor-nih mailing list
> Xplor-nih at nmr.cit.nih.gov
> http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih

Reply via email to