From: Ana Celia Vila Verde (acavilaverde_at_gmail.com)
Date: Thu Sep 26 2013 - 03:22:47 CDT
Hi Eddie,
You might want to look at these two approaches to model gold, both of
which are classical and should work with NAMD. In the first approach
(see papers by Ana Vila Verde and Janna Maranas,
http://pubs.acs.org/doi/abs/10.1021/bm9002464 and
http://dx.doi.org/10.1021/la104814z) the gold model is a simple LJ
potential; this approach, while simplistic, works well if you are
interested in a system with many charges (like, say, water+ions+proteins
adsorbing on gold) because the image charge interactions largely cancel
out. If you choose to go this way, I have a script to produce a psf
file for gold that would be a useful starting point for you. The script
is clunky, though, so you might be better off using the topotools tool
like Axel said...
The second approach involves a classical representation of gold
polarizability (see papers by Stefano Corni and E. Iori,
http://onlinelibrary.wiley.com/doi/10.1002/jcc.21165/abstract and
http://pubs.acs.org/doi/abs/10.1021/la904765u). You can try emailing
the authors to see if they have any files/scripts to run simulations in
NAMD before you start reinventing wheels in this case. I met some of
the authors and they are very approachable.
I hope this helps. Best,
Ana
On 9/25/13 7:37 PM, Axel Kohlmeyer wrote:
> On Wed, Sep 25, 2013 at 6:52 PM, Dr. Eddie <eackad_at_gmail.com> wrote:
>> I thought the protein would be the hard part...
>>
>> In the end I would like to look at effect the nanoparticles have on ordering
>> the system. Thus, to the best I can see right now, I would want the
>> course-grain model to include a local and global approximation to the
>> nanoparticle's electric field and surface polarizability due to water and
>> the proteins. I only expect the gold nanoparticle to interact via high order
>> electric moments, that depend on its shape, with the proteins. Is that way
>> off the mark?
> i don't know. i am not an expert in that field. it looks to me like a
> multi-scale problem, where you need more than just one kind of
> calculation, but need to build a coarser scale model based on the
> previous level and - if available - validate it with suitable
> experimental data. you mention polarization, but i would expect that
> this would be mostly determined by the surrounding solvation shell and
> attached molecules, so there is a good chance that this could be very
> well approximated with a non-polarizable all-atom model. there is the
> group of gary grest at sandia that does pretty big simulations related
> to that. perhaps, there is something you can learn from their
> publications.
>
> not sure how you can coarse grain this efficiently. and specifically
> include electrostatics well enough. it is possible to fit a multi-pole
> expansion to a set of point charges.
>
> at the higher level, you probably need something that is purely shape
> based and has an efficient to compute solvent. martini style coarse
> graining may be not enough for a reasonably large system, but could be
> used at an intermediate step.
>
> at the high level, it looks like you want to use something that models
> particles purely based on shape, but with a choice of shape variation.
> that would require either an implicit solvent via brownian dynamics or
> you could look at stochastic rotation dynamics (SRD). the group of
> sharon glotzer does work on "shaped nanoparticles", so perhaps you can
> steal some good ideas from them.
>
> there are likely many more people working on aspects of these kind of
> systems. i doubt that you will get something useful by just setting up
> a system with "some model" and let it go. this rarely works. MD
> simulations almost always need "a plan(tm)".
>
> axel.
>
>> In the end I will be trying to replicate a colleagues experimental results
>> of order inducement as a function of gold nanopartcile concentration and
>> other measures.
>>
>> Thanks again!
>> Eddie
>>
>>
>> On Wed, Sep 25, 2013 at 11:34 AM, Axel Kohlmeyer <akohlmey_at_gmail.com> wrote:
>>> On Wed, Sep 25, 2013 at 6:25 PM, Dr. Eddie <eackad_at_gmail.com> wrote:
>>>> Hi all,
>>>> I would like to use namd to perform a simulation with coarse-grain
>>>> proteins
>>>> around gold nanoparticles. I see vmd has a coarse grain model builder so
>>>> that leaves the gold nanopartciles. Are there any tools for building
>>>> these
>>>> nanoparticles and generating their psf using the charmm forcefield? I
>>> do you have an idea how you want to model / coarse grain the gold
>>> atoms? and how the interaction between the proteins and the gold
>>> particle? just having a tool to build the coarse grain model for the
>>> proteins, is the least of your problems.
>>>
>>> building custom psf files from scripting can be done with the topotools
>>> plugin.
>>>
>>> axel.
>>>
>>>> thought I'd ask before reinventing the wheel so to speak.
>>>> Thanks!
>>>> Eddie
>>>
>>>
>>> --
>>> Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0
>>> International Centre for Theoretical Physics, Trieste. Italy.
>>
>>
>>
>> --
>> Eddie
>
>
This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:43 CST