Re: [AMBER] question about scmask

From: <steinbrt.rci.rutgers.edu>
Date: Fri, 6 May 2011 03:28:57 -0400 (EDT)

Hi,

are the total free energy changes different? The result from one
transformation substep cannot easily be compared to that of some other
one, only results from complete thermodynamic cycles are physically
interpretable. For example, since you have removed different charge sets
in the first step of the two cases, you will get different distributions
of water around the molecules for the vdw steps, leading to different
deltaGs (different, but not necessarily wrong in either case).

If doing the complete transformation twice along different routes (i.e.
with different scmasks) gives very different results, then something is
indeed wrong, but you cannot know that just from partial results of the
substeps.

That said, if your substep results differ enourmously, by say 100kcal/mol,
then something is clearly wrong...

Thomas

On Thu, May 5, 2011 8:53 pm, Lishan Yao wrote:
> Hi,
>
> Thank you for the response. I think what makes me confused is the scmask
> keyword. In the first case,
> scmask=':3.CB,HB,CG2,HG21,HG22,HG23,OG1,HG1' for GTG and
> scmask=':3.CB,HB1,HB2,HB3' for GAG
> my understanding is that the atoms including
> ':3.CB,HB,CG2,HG21,HG22,HG23,OG1,HG1' are decoupled with softcore method
> from the surroundings while the atoms including scmask=':3.CB,HB1,HB2,HB3'
> are coupled to the surroundings.
> In the second case,
> scmask=':3.HB,CG2,HG21,HG22,HG23,OG1,HG1' for GTG and
> scmask=':3.HB1,HB2,HB3' for GAG
> atoms including ':3.HB,CG2,HG21,HG22,HG23,OG1,HG1' are decoupled with
> softcore method from the surroundings while the atoms including
> scmask=':3.HB1,HB2,HB3' are coupled to the surroundings. If my
> understanding is correct, I would expect the same deltaG values from one
> transformation since the initial and final state are the same for both
> cases (CB is a common atom in GTG and GAG, I would think it doesn't matter
> if I exclude it from scmask).
>
> Best,
> Lishan
>
>
>
> ----- Original Message -----
> From: <steinbrt.rci.rutgers.edu>
> To: "AMBER Mailing List" <amber.ambermd.org>
> Sent: Thursday, May 05, 2011 6:33 PM
> Subject: Re: [AMBER] question about scmask
>
>
>> Hi,
>>
>> two things:
>>
>> a) Are you talking about deltaG values from one transformation? These
>> will
>> certainly be different. Only deltadeltaG values from a properly set up
>> thermodynamic cycle will give interpretable results.
>>
>> b) If you do get very different ddG values, this indicates that your
>> simulation is not converged. You don't give simulation details, so it's
>> impossible to judge that.
>>
>> Kind Regards,
>>
>> Thomas
>>
>> On Thu, May 5, 2011 5:29 am, Lishan Yao wrote:
>>> Dear Amber users,
>>>
>>> I set up a alchemy simulation for the GTG to GAG mutation in water
>>> (with head and tail groups neutralized). Following the typical
>>> procedure, I did electrostatics first and then vdw. In the vdw alchemy
>>> process, I used softcore method. But I got very different dv/dl values
>>> when I use
>>>
>>> scmask=':3.CB,HB,CG2,HG21,HG22,HG23,OG1,HG1' for GTG and
>>> scmask=':3.CB,HB1,HB2,HB3' for GAG
>>> or
>>> scmask=':3.HB,CG2,HG21,HG22,HG23,OG1,HG1' for GTG and
>>> scmask=':3.HB1,HB2,HB3' for GAG
>>>
>>> I can't figure out what is the problem. It seems to me they should give
>>> same or at least similar results.
>>>
>>> Best,
>>> Lishan
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>
>>
>>
>> Dr. Thomas Steinbrecher
>> formerly at the
>> BioMaps Institute
>> Rutgers University
>> 610 Taylor Rd.
>> Piscataway, NJ 08854
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>


Dr. Thomas Steinbrecher
formerly at the
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri May 06 2011 - 01:00:03 PDT
Custom Search