About a REMD tutorial

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Sat Sep 07 2013 - 09:35:51 CDT

It is my experience namd has given little guidance as to temperature remd.
I may well be wrong, especially as the temperature distribution is
generated automatically. However, besides attempting by trial and error
(not mentioned the rule of thumb of number of replicas = square root of
total number of atoms, if valid for namd too) to get exchange by changing
the number of replicas in a very restricted range of temperatures, there is
little help about how to set the other parameters to get proper relaxation
at each temperature. I was also unable to find how to restrain the aa
chirality, which may well be needed if 600K are touched. Such problems may
be irrelevant for the typical tests presented, with peptides of a few
atoms. like deca-alanine in implicit medium, but they arise for large
systems, typically for systems in explicit solvent. With these, the limits
of current instrumentation are easily reached, not to say about the cost,
so that one tends to a perfect balance before putting money in.

I can only suppose that I failed to come across the right literature, or
the namd user is supposed to be particularly smart.

Thanks for pointing to technicalities for temperature remd. Published
papers are mostly with amber, or give little technical advice for the flags
offered by namd

francesco pietra

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:41 CST