AMBER Archive (2006)

Subject: Re: AMBER: Replica exchange

From: Carlos Simmerling (carlos_at_csb.sunysb.edu)
Date: Mon Jul 17 2006 - 09:44:55 CDT


also note that REMD in amber requires that inpcrd files
have velocities, so you have to run MD for at least a
few steps to get velocities. You can do that for each replica or use
a single restart file with velocities as inpcrd for all replicas.

Guanglei Cui wrote:

> Hi Austin,
>
> I think it's better to have the replicas separately equilibrated to
> their target temperatures before you commit them to the exchange. Is
> there any particular reason that you want to cold start all of them
> other than convenience?
>
> Guanglei
>
> Austin B. Yongye wrote:
>
>> Dear amber users,
>> I am running some replica exchange simulations and split between two
>> options:
>>
>> Is it better to
>> 1. Equilibrate a single structure in a solvent to a particular
>> temperature
>> eg 300K and then generate the replicas from this point with different
>> target temperatures, irest=1, ntx=5, etc?
>>
>> 2. Start with all the replicas at 0K and give each replica a different
>> target temperature, irest=0, ntx=1, etc?
>>
>> Thanks,
>> austin-
>>
>> -----------------------------------------------------------------------
>> The AMBER Mail Reflector
>> To post, send mail to amber_at_scripps.edu
>> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu

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