AMBER Archive (2009)

Subject: Re: [AMBER] minimization to reduce covalent deviation

From: David A. Case (case_at_biomaps.rutgers.edu)
Date: Tue Jun 02 2009 - 08:38:58 CDT

On Mon, Jun 01, 2009, Peter Varnai wrote:
>
> I tried conjugate gradient minimisation in explicit water environment
> using sander in amber10 (no shake but the same pbc used in md) and
> also tried stripping the water and using GB to reduce the degree
> freedom. I can get the rms of the gradient below 0.01 kcal/mol/A after
> a couple of thousand steps but not much lower. I would have thought
> that this should be enough to be within say 0.015 A and 2 degrees from
> the ff values (which are somewhat different from dictionary standard
> used for nucleic acids by PDB). However, I get a long list of covalent
> violations mostly for angles (about 4 degrees) and base planarity. Eg,
> N3-C2-O2 in CYT is 116.2, equilibrium value in parm99.dat is 122.5 and
> the standard value in PDB is 121.9.

I don't think this has anything to do with the quality of the minimizer.
I see small N3-C2-O2 angles in cytosines even at very small RMS gradient
values. It looks like it is making this distortion in order to get a better
H-bond to the guanine base it is paired with (that's just a speculation).
You could see what you get with a single-stranded molecule as a test. But, on
initial inspection, this looks like a force field issue, not a program issue.

...dac

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber