AMBER Archive (2000)

Subject: Re: problems in sander

From: Yong Duan (yduan_at_heidrun.compchem.ucsf.edu)
Date: Sun May 07 2000 - 15:11:36 CDT


Dear Devi,

Here are my wild guesses.

1) dynamics
The system was at 300K at 20.000ps and you restarted the simulation. It
jumped to 426K and came back immediately. My guesses is that you
reassigned the velocities.
If so, nothing needs to be worried. The velocity reassignment module
reassigns velocities to all 3N dimensions when actually you have 3N-m
degrees of freedom (m is about N because of SHAKE). Those velocity
components are immediately removed within one or two steps and the
temperature goes back to normal. However, if you are really concerned,
re-do it without reassigning velocities.

2) minimization
My guess is that your DISO (or DISH) was small in your BOX option. If so,
that can cause water molecules in close contact with ions. Yet, since the
distances between ions and phosphate oxygens were initially 5.0A, it is
expected that some of the gaps may not be filled by water. Pressured by
the close contacts of water, these ions may move toward phosphates. If
these are true, you may do the following to circumvent the problem.
a) Build the system with larger DISO (I am comfortible with 2.8A). Or
build the system with Leap, which is more user-friendly than EDIT.
b) A short minimization. 23K steps of minimization, when done in small
system (i.e., tens of atoms), can bring you very close to the true local
energy minimum, but only local energy minimum. Given that you are dealing
with macromolecular system, it will only bring you close to local energy
minimum (regardless how hard you minimize it). Since the goal is to
remove bad contacts and to have a reasonable starting structure for
dynamics, I typically use 200-500 steps energy minimization. In my
opinion, anything beyond 500 steps is a waste of efforts. Your objective
can be better served by running a short MD at low temperature which
can usually lower energy a lot more efficiently than energy minimization
(because of the rugged energy landscape, minimization = MD at 0K,
making the system very glassy) since 23K is equivalent to about 46 ps.
c) good luck!

yong

Department of Pharmaceutical Chemistry Phone: (415) 476-8291
University of California Fax: (415) 502-1411
San Francisco, CA 94143

On Sun, 7 May 2000, devi rvel wrote:

> Hello,
> Problem 1:-dynamics
> I got some problem while doing dynamics for my 14mer dna molecule(
> using PME ) . During my production run I heated the molecule to 300K for
> 20 ps and for rest of the cycles I wanted to maintain the temperature at
> 300k itself. This I split it into two parts as follows:
> 1) heating from 10k to 300k for 20ps
> 2) maintaining the temperature at 300k for 700ps.
> When I started my second part of the dynamics, using the output
> co-ordinates of the first step, I got intially 426K ,eventhough it is
> heated upto 300k in the previous step.Immediately after 0.004ps (ie
> 20.004ps) it is coming to 300k and stays there itself for the rest of the
> cycles. I am not able to find out the reason for this. If there is any
> other way to rectify this problem, please let me know about it.
>
> Problem 2:-minimization
> I had problems in fixing ions for my dna molecule using the “edit”.
> Initially I fixed the ions at the bisector of O1P-P-O2P angle and 5A
> distance away from the phosphate and I used box option to add water
> molecules. There I got few ion-ion short contacts and I carried out my free
> minimization(with PME) using the same co-ordinates itself. After 23141
> cycles the molecule reaches the RMS gradient of 0.01kcal/mol/A . But few
> ions are moving close to the phosphate and anionic oxygens to a distance of
> 3.5A and 2.3A respecitvely.Some ion-ion contact is also there.
>
>
> Please , help me to solve these problems .
>
> devi
>
> ________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
>