Hi,
> Hello. While doing TI simulations, what should the softcore selection be.
> What I mean is, will I get different results if I choose only the atoms
> that are changing or if I choose the residues that contain the atoms that
> are changing? I theory I would think that this should not matter, since
> DV/DL is calculated for all the differences in both systems, and all atoms
> that are the same contribute in nothing to DV/DL. Is that the correct way
> to think, or am I missing something here?
You are right, in theory you will obtain the same overall result (but DVDL
for substeps will differ) but in practice they will differ due to
different sampling efficiency. Adding more atoms to the SC region means
more of a dual topology model, giving your system greater freedom of
motion. If this will speed up or slow down sampling I'm not sure and that
may well be system dependent.
So you can put as few atoms into SC as possible, or decide to put
reasonable chemical fragments into SC, or entire residues, etc. None of
this is wrong, but may lead to very different convergence behaviour.
Thomas
Dr. Thomas Steinbrecher
formerly at the
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Mar 01 2013 - 01:00:02 PST
