AMBER Archive (2009)

Subject: Re: [AMBER] implicit and explicit simulation: salt effect

From: David LeBard (david.lebard_at_asu.edu)
Date: Fri Jan 09 2009 - 22:23:24 CST

As far as I know, there is no parameter describing the Debye screening
length for ions added to explicit solvent simulations. From this
perspective, the effect of adding an ion (or increasing the salt
concentration) to a real solution is mostly missing in explicit MD
simulations because an MD ion cannot screen the electric field arising
from the other charges in the simulation due to the lack of a Debye
length parameter. Therefore, it is my understanding that MD ions in
explicit solvent simulations are simply there to create the necessary
zero net system charge needed for the Ewald summation.

That said, I would love to hear what the real experts have to say
regarding this interpretation, because it has been a point of
confusion for myself for the past few years.

Good luck,
David

On Fri, Jan 9, 2009 at 10:49 PM, David Watson <dewatson_at_olemiss.edu> wrote:
> On Jan 9, 2009, at 9:03 PM, Catein Catherine wrote:
>
>>
>> Dear Sir/Madam, I would like to study the salt effect for the binding of a
>> neutral drug with DNA from 0M to 0.5 M of salt in a QUALITATIVE manner. It
>> seems to me that I can do the GB simulation with salt effect considered.
>> According to amber website, it seems to me that it is not a problem at all
>> to do the simulation up to 0.2M salt. What about up to 0.5 M of salt for
>> qualitative analysis? If I should only do the analysis with explicit
>> solvent, how can I incorporate salt effects in the simulation? what is the
>> command line should I used? Best regards, Cat
>
> As to performing GB with the modified model, I would suggest that you read
> the original article [Theor Chem Acc (1999) 101:426-434] and note that the
> model was developed (as far as I can tell) with the GBHCT implementation.
> Several other GB implementations have been developed over the years and I
> would be reluctant to use the AMBER saltcon parameter with the GBn model,
> based on personal experience.
> Your mileage may vary.
>
> As to incorporation of higher concentrations, the original article mentions
> that the mathematical model comes from the Debye-Hueckel limiting law, in
> which case I would also be skeptical of your intentions with regards to salt
> concentration as 0.5 M should be well above the limits where that formalism
> holds.
>
> With regard to explicit solvation, you could could first neutralize your
> system with monovalent counterions in tleap/xleap, in which case I would
> recommend that you look at the first tutorial in the series on the
> ambermd.org website, which mentions a DNA system, and discusses the
> influence of the phosphate in the DNA backbone on charge considerations.
> Then you would have to increase/decrease the size of your periodic box in
> order to adjust the concentration of ions in your system.
>
> During the setup of your system, you would have to determine the mean volume
> that your system will eventually occupy, which would be difficult in a
> periodic boundary simulation. Then you would need to determine what you mean
> by "0.5 M salt", because monovalent salts will have different equivalency
> than salts such as CaCl2 or KH2PO4... the sizes of said ions would influence
> your decisions as well, not to mention trying to parameterize some of the
> ions in order to be compatible with Amber force fields.
>
> This is an interesting and enlightening question, and I can't wait to hear
> from Case and Simmerling, because I could use some education on this matter,
> as well.
>
>
> _______________________________________________
> 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