AMBER Archive (2009)

Subject: Re: [AMBER] help regarding step size

From: Carlos Simmerling (carlos.simmerling_at_gmail.com)
Date: Wed Apr 22 2009 - 06:03:28 CDT

in my opinion nothing in the solute should be allowed to move before
all of the water and anything else added (H atoms, missing side
chains, ions, etc) has fully equilibrated. otherwise, the way you add
water affects the solute.minimizing is not enough, you must do NPT MD
until pressure/volume, energy etc have stabilized.

it's important to note that just because things look good later
doesn't mean the equilibration worked. you may have solute structural
changes and then volume, density etc all look fine.

once you do this (and again I recommend doing it over with a good
equilibration protocol) then you can switch to 2fs steps. don't rely
on what you have already done, I would not consider it to be
acceptable if I were reviewing a paper.

On Wed, Apr 22, 2009 at 6:50 AM, Vikas Sharma
<vs_vikassharma_at_yahoo.co.in> wrote:
> Hi carlos...
> thanks for the reply..
> I first minimized the hydrogens and water, the minimized the whole system, then heated the sytem for 50 ps(50000 steps step size 1fs), then equilibrated for 100ps(1000000 steps,step size ifs)..
> my questions are:
> 1. do i need to equilibrate more(the density has stabilized to 1.03 and temp has also stabilized around 300)
> 2. Now for production MD ca i change the step size from 1fs to 2 fs?
>
>
>
>
> ________________________________
> From: Carlos Simmerling <carlos.simmerling_at_gmail.com>
> To: AMBER Mailing List <amber_at_ambermd.org>
> Sent: Wednesday, 22 April, 2009 3:43:54 PM
> Subject: Re: [AMBER] help regarding step size
>
> do you mean you did a single heating run and that's all for equilibration?
> probably not nearly good enough. a good equilibration protocol should include
> multiple steps, with slow turning off of restraints on various sections.
> first equilibrate the solvent fully, along with H atoms and anything
> else added. keep
> what you know from expt restrained. minimize first, then do heating
> MD. then, slowly turn
> off restraints on side chains, since they can be involved in surface
> packing or may otherwise need to move.
> only then do you turn off restraints on the backbone. similar ideas
> should be applied to non-protein.
> also we use molprobity first to check for Asn/Gln flips in the initial
> structure.
> if you just build, head and go I'd expect some real problems even if
> it doesn't crash.
>
>
>
> On Wed, Apr 22, 2009 at 4:57 AM, Vikas Sharma
> <vs_vikassharma_at_yahoo.co.in> wrote:
>> hi all..
>> I am using MM PBSA..
>>
>> the thing is that i first heated the protein for 50000 steps with 1 fs time step(50 ps total)
>>  coz when i used 25000 steps with 2fs time steps there was error regarding velocities...
>>
>> so i used 50000 steps with 1fs time step for heating...then i equilibrated for double the no. of steps i.e. 100000(100ps total) in heating steps with 1fs time step...(hope this much equilibration would be ok as in tutorial its been asked to equilibrate fr 250 ps....)
>>
>> now i want to run production MD for 1ns using 2fs time step..wont there be any problem coz i will be varying the step size from 1fs (during heating and equilibration) to 2fs during production MD..
>>
>> thank you
>>
>>
>>
>>      Bollywood news, movie reviews, film trailers and more! Go to http://in.movies.yahoo.com/
>> _______________________________________________
>> AMBER mailing list
>> AMBER_at_ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
>      Now surf faster and smarter ! Check out the new Firefox 3 - Yahoo! Edition http://downloads.yahoo.com/in/firefox/
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber