On Thu, Nov 27, 2014 at 12:40 PM, Mary Varughese <maryvj1985.gmail.com>
wrote:
> ok sir,
> then it would be better to allow the structure to evolve itself, adjusting
> torsions and helical. isn't?
>
That's what I would do.
> Sir, could you suggest me any criteria in arriving at a reliable structure?
>
If you have experimental measurements, compare against those. Otherwise,
just check that your structure is converged (e.g., that average structures
over the first and second halves of the simulation are the same, for
example).
> may be rmsd convergence?
RMSD isn't always the best for determining whether a simulation has
converged (at times it can tell you when one _hasn't_).
>
> would you prefer long simulation or multiple in
>
> such a case?
>
Multiple long simulations :). This is a loaded question with differing
opinions, and it often depends on the system of interest (e.g., are there
any long time-scale motions?). If you can afford both (and/or construct
"longer" simulations from shorter ones using techniques like Markov state
models), I would suggest that approach.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Nov 28 2014 - 19:30:03 PST
