[gmx-users] A question on REMD result analysis

[Christopher Neale]

Most often, people only analyze conformations drawn from the target temperature (e.g. room temperature). 
There are reasons to look at higher temperatures (e.g. to decompose entropy/enthalpy or if your system has 
biologically or chemically relevant temperature-dependence) but you can probably ignore that for your initial 
analysis.

You should not eliminate any conformations drawn immediately after an exchange. The whole point of 
REMD is that all of your conformations are in thermal equilibrium, even those right after an exchange. 
However, you should do some block averaging and consider discarding some of the data obtained from the 
initial portion of you REMD simulation (just as you would for a standard simulation).

Chris.

-- original message --

I’m trying to run a REMD simulation on a protein. After a lot of reading,
I am still unclear on the result analysis. Is it correct that, of many
replicas, I should select the frames from only the replica with my desired
temperature?(ex, room temp.) If not correct, how do I select the frames
from a REMD simulation as my final structures? If correct, should I ignore
some frames right after an exchange has occurred? I think they would not
have reached equilibrium and therefore not to be used.

I am quite confused. Any replies would be of great help.

Many thanks in advance,