If you are certain that your sampling is sufficient for your
oligosaccharide, then the results you get, in some sense, reflect what
the force field will do. But, you might still need better sampling.
In nature, the addition of salt lowers the dielectric constant of the
water, decreasing the extent to which solvent-separated charges affect
each other. However, in a typical MD simulation like the one you
describe, there are no extended effects of the ions on the water or
the solute (e.g., polarizability). So, the only time the ion would
have an effect is when it is more-or-less directly associated with the
glycosidic linkage. I just took a look at an MD of sucrose I happen
to have, and there isn't a lot of space for water between the atoms
relevant to the glycosidic linkage. Of course, a 1-6 linkage will be
a little roomier, if you have one of those. My point is that within
this sort of simulation, you might need to ensure that you
sufficiently sample interactions of the ions with the glycosidic
linkage. You could monitor this by tracking the locations of the ions
with respect to the linkage. You might also try increasing the
non-bonded cutoff distance so that the ions' effects are counted more
often. This latter will slow your simulation, of course. And, in the
end, you may still find that the lack of polarization is the biggest
effect.
There have been simulations published recently of carbohydrates in
ionic liquids. Have you looked at those? They might help you to find
relevant experimental or computational data for comparison.
:-) Lachele
On Tue, Oct 5, 2010 at 11:07 AM, Sergey Samsonov
<sergeys.biotec.tu-dresden.de> wrote:
> Thank you, Lachele!
>
> On 10/05/2010 03:36 PM, Lachele Foley (Lists) wrote:
>> What sort of sampling method did you use? What water model? If you
>> did not sample sufficiently, then you simply aren't observing enough
>> simulation to tell. How did you sample (REMD, long simulation, etc.)?
> I use 20ns MD in a box of TIP3 waters (10 A from the sugar to the box
> border in xleap). It looks like I sampled sufficiently, at least for
> dihedrals data.
>> Whenever possible, you should validate any modeling method against
>> experimental data that correlates to the sort of simulation you are
>> doing. Are there no experimental studies at all to use for
>> comparison?
> Nothing in respect to ionic strength but the data to compare for
> sampling is available.
>> It is possible that the behavior you see is indicative of natural
>> behavior, though I doubt it is a perfect reflection. If there are
>> deviations, I would expect them to be related to the absence of
>> polarizability in the model.
> Exactly, I also expected that since it could have been more polarized,
> the charged parts of the molecules would behave a bit differently but it
> doesn't look like this case.
>
> Cheers,
>
> Sergey
>> :-) Lachele
>>
>> On Tue, Oct 5, 2010 at 6:00 AM, Sergey Samsonov
>> <sergeys.biotec.tu-dresden.de> wrote:
>>> Dear All,
>>>
>>> I've been checking, how the flexibility of some saccharides changes (in
>>> terms of glycosidic linkages dihedrals) depending on ionic strength.
>>> What I found is that there are neither quantitative, no qualitative
>>> differences for the effect of ionic strength in the range I used (from 0
>>> up to quite extreme 5M for NaCl). Do you think this could be explained
>>> by real physical nature of sugars flexibility or it is rather the force
>>> field (I used ff03 + GLYCAM06) inability to find such effects as changes
>>> in flexibility of molecules (or their diffusion constant, which I also
>>> checked) under different values of ionic strength in solvent?
>>>
>>> Thanks in advance and best regards,
>>>
>>> Sergey
>>>
>>>
>>>
>>> _______________________________________________
>>> 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
>
--
:-) Lachele
Lachele Foley
CCRC/UGA
Athens, GA USA
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Oct 05 2010 - 12:00:04 PDT
