[gmx-users] Replica exchange temp. distribution

[Mark Abraham]

sarbani chattopadhyay wrote:
>  
> Hi everyone,
>                   I will like to do Replica exchange simulation on a 
> peptide. We have a single
> machine with 4 cpus and the gromacs3.3.2 is installed.
> Thus I will be able to select only 4 temperature values. I came across 
> the  "T-REMD" caculator
> for temperature distributions . However over a high range of temperature 
> there are more than
> 4, values given in the optimal tempearture distribution as obtained as 
> an output from the "T-
> REMD" calculator. 
> How to choose the best possible 4 temperature values.

Using explicit solvent with a low number of replicas will severely limit 
the temperature space you can span while achieving exchanges in 
accessible simulation times. See http://dx.doi.org/10.1021/ct800016r and 
references therein.

Mark