May I please ask what you mean by sample?
Ramin
Get Outlook for Android<
https://aka.ms/ghei36>
________________________________
From: David A Case <david.case.rutgers.edu>
Sent: Thursday, July 20, 2017 10:26:54 AM
To: Ramin Salimi
Subject: Re: [AMBER] problem Finding entropy
On Thu, Jul 20, 2017, Ramin Salimi wrote:
>
> As the final question, you said 4ns with 2000 snapshots shouldn't give
> convergence for entropy. why is that so? If the output file shows the
> energy converges by the end of the run, wouldn't that be enough? I mean,
> is convergence criterion any different for the entropy?
There are two questions here: For a given length of simulation (say 4ns), you
still need lots of samples in order to construct an accurate covariance
matrix. You can't even get a full-rank covariance matrix with fewer than
3*natom samples, and you generally need lots more than that. Some
computational experiments with *your* system will provide much more
information.
The second point is whether 4 ns of simulation is enough to sample
configurational space. Again it depends on your system, but 4 ns seems
pretty short to me. You seem to be expecting "bubbles" (single-stranded
regions?) that you don't see. Remember that none of the configurational
entropy estimates will be very accurate, so concentrate your efforts
on getting good simulations that make sense, and which provide useful
information, before worrying too much about the entropy numbers.
....dac
--
====================================================================
David A. Case | david.case.rutgers.edu
Dept. of Chemistry & Chemical Biology |
Rutgers University | office: +1-848-445-5885
174 Frelinghuysen Road, Rm. 208b | cell: +1-609-751-8668
Piscataway, NJ 08854 USA | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fcasegroup.rutgers.edu&data=02%7C01%7Cramin.salimi01%40utrgv.edu%7C9a88981ee6f74e86e18c08d4cf948469%7C990436a687df491c91249afa91f88827%7C0%7C0%7C636361684164267312&sdata=GXikt12lXfczYlfwJ8zFxjcftFrGIlTQv7T2nzLFFKM%3D&reserved=0
====================================================================
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jul 20 2017 - 11:00:05 PDT
