AMBER Archive (2005)

Subject: Re: AMBER: Replica exchange

From: Carlos Simmerling (carlos_at_ilion.bio.sunysb.edu)
Date: Sun Aug 28 2005 - 08:16:46 CDT

keep in mind that the number of replicas scales with the
square root of system size. A peptide with implicit solvent
may need 10-30 replicas, while explicit solvent in a box large
enough to unfold may need hundreds of replicas. Read the literature
to see what people have done. Implicit solvents are faster and much
cheaper for REMD but they do have limitations. Choosing between
implicit and explicit solvent (especially for REMD) is
a difficult decision. Amber9 will have some new REMD features that
will help with this, but REMD in explicit solvent for all but the
smallest peptides is not practical unless you have huge computers
available.

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu