Hi all,
I believe we found the problem ...
Here it is:
In the "ions94.lib" file (from the AMBER distribution) that we were
loading for the ions, the "K+" leap unit is assigned a "K" type ...
Therefore every time a "K+" is added in Leap it is directly assigned to
a type "K" ..
Our frcmod was actually modifying the "K+" type ...
If I modify the "ions94.lib" and replaced the "K" with "K+" everywhere,
the K+ is now properly overwritten from the frcmod.
This is a very easy-to-make mistake so I hope this will help other
people when they try to update K+ parameters via frcmod files.
Maybe the developers could consider clearing up the different libraries
so that such mistakes are avoided ....
Best,
Vlad
On 01/25/2013 01:49 PM, Vlad Cojocaru wrote:
> Hi Tom,
>
> We did some careful checking and indeed you were right .. For some
> reason the frcmod we were using for the Smith and Dang parameters (see
> below) was not overwriting the Aqvist parameters from the parm99.dat
> ... Weird is that this is only happening for K+, not for Na+, not for
> Cl- (these are correctly overwritten)....
>
> Below is the frcmod we are using .. If you or anybody else has any
> idea what could be the problem that this frcmod is not overwriting the
> original parm99.dat, please give us a hint ...
>
> Best wishes
> Vlad
>
> -------frcmod.ions_dang
> smith/dang ion parameters
> MASS
> IM 35.45 3.235 assumed to be Cl- (ion minus)
> IP 22.99 0.250 assumed to be Na+ (ion plus)
> Na+ 22.99 0.250 assumed to be Na+ (ion plus)
> Cl- 35.45 3.235 assumed to be Cl- (ion minus)
> K+ 39.10 1.060 assumed to be K+ (ion plus)
>
>
> NONB
> IM 2.47 0.1 Smith/Dang
> IP 1.32 0.13 Smith/Dang
> Na+ 1.319 0.13 Smith/Dang
> Cl- 2.47 0.1 Smith/Dang
> K+ 1.869 0.1 Smith/Dang
>
> On 01/24/2013 09:43 AM, Vlad Cojocaru wrote:
>> Dear Tom,
>>
>> Thanks a lot for this detailed answer.
>> To clarify, when talking about Dang parameters, I actually referred to
>> Smith/Dang, in fact exactly the parameters you wrote. You actually sent
>> me the parameters a while ago and I also double-checked the original
>> paper ...
>>
>> One difference is that we are using TIP3P .... Not sure if this explains
>> things but maybe we could do a control simulations using SPC-E to check
>> if we see the same ...
>> I will also double check with Felipe (the PhD student working with me)
>> because he actually observed these clusters ...
>>
>> I'll keep you and the amber list posted
>>
>> Best wishes
>> Vlad
>>
>>
>> On 01/24/2013 01:41 AM, Thomas Cheatham wrote:
>>>> Recently we are observing the formation of KCl crystals in simulations
>>>> of about 100 ns with 100 mM KCl. We are using KCl to reproduce the
>>>> experiments we are trying to compare with. For these simulations, we
>>>> used the Dang parameters for ions as those were reported to not
>>>> lead to
>>>> the formation of salt crystals ?
>>> Are you sure you are running with what you think you are? Dang
>>> parameters
>>> with Aqvist K+ lead to immediate crystal formation. I do not know
>>> about
>>> Dang K+/Cl- (or if there *is* a Dang K+). However, Smith + Dang
>>> appear to
>>> be fine as has been reported in the ABC simulations and I just
>>> looked for
>>> one of my long trajectories of DNA with KCl Smith/Dang w/
>>> Lorentz-Bertholet (standard AMBER) combining rules... Parameters are
>>> (SPC/E water):
>>>
>>> Smith / Dang:
>>>
>>> RDPARM MENU: printty
>>> NOTE: if either A or C is zero (*), we cannot infer
>>> the value of r* or epsilon and assume zero...
>>>
>>> Type r* eps
>>> * HO 0.0000 0.0000
>>> OH 1.7210 0.2104
>>> K+ 1.8687 0.1000
>>> Cl- 2.4700 0.1000
>>>
>>> In simulations out to 10 microseconds of DNA in higher KCl salt we
>>> do not
>>> see crystal formation and only transient ion-pairing (very low
>>> occupancy
>>> ~0.25% and short lifetimes). This can be seen in radial distribution
>>> functions attached. Top is default/old AMBER Aqvist K+ and Dang Cl-
>>> which
>>> immediately forms crystal. Below is the RDF from my DNA simulation...
>>>
>>> --tec3
>>>
>>> (I haven't checked S+D in TIP3P since we tend to run with Joung
>>> parameters
>>> except for the ABC consortium simulations which use S+D since they were
>>> run previously; however in TIP4PEW I do not think we see crystals
>>> either
>>> with S+D but I will check).
>>>
>>>
>>>
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Dr. Vlad Cojocaru
Max Planck Institute for Molecular Biomedicine
Department of Cell and Developmental Biology
Röntgenstrasse 20, 48149 Münster, Germany
Tel: +49-251-70365-324; Fax: +49-251-70365-399
Email: vlad.cojocaru[at]mpi-muenster.mpg.de
http://www.mpi-muenster.mpg.de/research/teams/groups/rgcojocaru
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Jan 25 2013 - 05:30:04 PST
