From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Tue Nov 04 2014 - 11:39:49 CST

On Tue, Nov 4, 2014 at 11:58 AM, Nifeng Guo hui <nifenggaohui_at_gmail.com> wrote:
> Hey,
> After I paly with RDF Plugin, I am still unsure about how properly use it.
> First, "it suddenly becomes zero, because you run out of data. under
> consideration of minimum image conventions you usually are only able to
> compute the g(r) up to half the shortest box length. " Does this mean it is
> not appropriate to treat all the 36 amino acids as one whole. Second, if my
> conclusion that it is not appropriate to treat all the 36 amino acids an a
> ball is validated, then does this explains why the RDF always increase from
> starting until a suddenly dropped point as shown in my first plot. Third,
> then what is the feature of "Use PBC" in the the Plugin?

please learn how a g(r) is computed and then we can discuss details.
your questions make little sense as they are.

> Thanks.
> Peng
>
> 2014-11-03 15:49 GMT-06:00 Axel Kohlmeyer <akohlmey_at_gmail.com>:
>
>> On Mon, Nov 3, 2014 at 4:29 PM, Nifeng Guo hui <nifenggaohui_at_gmail.com>
>> wrote:
>> > Hello, Axel and Tim
>> >
>> > I analyze RDF in production run of 10 ns. (After energy minimization and
>> > NVT, NPT Equilibration simulation.) In my understanding of RDF, after
>> > certain distance from the reference point, the particles will distribute
>> > evenly which means water molecules are not attracted or repelled by the
>> > protein molecule. Thus, the water density will appear as a horizontal
>> > line.
>> > However, it's hard to explain that RDF increases to 28 anstrom and
>> > suddenly
>> > becomes zero.
>>
>> it suddenly becomes zero, because you run out of data. under
>> consideration of minimum image conventions you usually are only able
>> to compute the g(r) up to half the shortest box length. the g(r)
>> calculation in VMD uses a smarter normalization function that can go
>> beyond that, with the maximum being half the length of the diagonal
>> through the box (with limited statistical accuracy, due to the
>> shrinking normalization volume).
>>
>> > Now, I change the code in "Section 2" as "name OH". The plot seems like
>> > understandable. So does this mean the code for "water" in "Section 2" is
>> > incorrect? I attach the new plot. Please take a look at it. Thanks.
>>
>> your question has nothing to do with the g(r) calculation but with
>> VMD's atom selection language. the g(r) plugin will compute what you
>> ask it to compute. whether you select the right group(s) of atoms for
>> your purpose is something that you have to validate independently.
>> when using macros like "water", you have to make certain, that your
>> input conforms to the underlying heuristics. this is not the g(r)'s
>> plugin's job, but yours. from the point of the plugin, there is no
>> "right" or "wrong". it just uses your input. please revert to the VMD
>> user's guide for details.
>>
>> axel.
>>
>>
>> >
>> > Peng
>> >
>> > 2014-11-03 0:58 GMT-06:00 Axel Kohlmeyer <akohlmey_at_gmail.com>:
>> >
>> >> On Sun, Nov 2, 2014 at 11:19 PM, Nifeng Guo hui
>> >> <nifenggaohui_at_gmail.com>
>> >> wrote:
>> >> > Hi Tim,
>> >> > Thanks for the explanation. Now I follow your instruction and plot it
>> >> > again.
>> >> > It still does not make sense. I use period boundary condition to
>> >> > simulate in
>> >>
>> >> well, the problem is very likely not with the g(r) plugin but with
>> >> your expectations. you most certainly don't have a homogeneous
>> >> monoatomic fluid as your system, which will lead to the graphs more
>> >> commonly seen as examples for a g(r). please review the definition of
>> >> the g(r) (i.e. what the graph means, not what it typically looks like)
>> >> and compare it to the structure you are feeding into VMD. i am very
>> >> confident that what you see as a graph is correct. it would be
>> >> consistent with a two component system that is mostly separated.
>> >>
>> >> > a cubic box with dimension 50. The molecule size is
>> >> > cellBasisVector1 48.12900161743164 0 0
>> >> > cellBasisVector2 0 42.979000091552734 0
>> >> > cellBasisVector3 0 0 43.9109992980957
>> >> > cellOrigin 0.5063787698745728 1.1608952283859253 0.11323882639408112.
>> >> > Please find the picture in the attachment. Thanks.
>> >> > Peng
>> >> >
>> >> > 2014-11-02 19:19 GMT-06:00 Tim Lo <timlo_at_hku.hk>:
>> >> >
>> >> >> Hi Peng,
>> >> >>
>> >> >> By looking at the r-axis, the graph just displays g(r) for the range
>> >> >> of
>> >> >> [0, 10]. You should increase "max r" to your desired value, but not
>> >> >> just use
>> >> >> the default value of 10.0.
>> >> >>
>> >> >> Tim
>> >> >>
>> >> >>
>> >> >> On 2014/11/3 上午 06:01, Nifeng Guo hui wrote:
>> >> >>>
>> >> >>> Dear VMD users,
>> >> >>> We perform one simulation with one protein immersed in water. Now,
>> >> >>> we
>> >> >>> want to analyze the water distribution around this molecule after
>> >> >>> finishing
>> >> >>> production run. The g(r) GUI Plugin, Version 1.3 was applied to
>> >> >>> calculate
>> >> >>> its function. After I input PSF and DCD files, I put "residue 1 to
>> >> >>> 36"
>> >> >>> in
>> >> >>> Section 1 and "water" in Section 2. (This protein is composed of 36
>> >> >>> amino
>> >> >>> acids. ) I also tried "protein" for Selection 1. Next, Frames are
>> >> >>> "0"
>> >> >>> in
>> >> >>> First and "19999" in Last. Then before using compute g(r), I select
>> >> >>> Use PBC,
>> >> >>> Display g(r), and Display int(g(r)). However, the result does not
>> >> >>> reach our
>> >> >>> predictions. The RDF and Coordination number are both increasing
>> >> >>> exponentially as r increase. In our prediction, it supposes to
>> >> >>> become
>> >> >>> constant. Could you help me find what cause this? Plese also take
>> >> >>> a
>> >> >>> look at
>> >> >>> the plot in the attachment. Thanks.
>> >> >>> Peng
>> >> >>>
>> >> >>>
>> >> >>>
>> >> >>
>> >> >
>> >>
>> >>
>> >>
>> >> --
>> >> Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0
>> >> College of Science & Technology, Temple University, Philadelphia PA,
>> >> USA
>> >> International Centre for Theoretical Physics, Trieste. Italy.
>> >
>> >
>>
>>
>>
>> --
>> Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0
>> College of Science & Technology, Temple University, Philadelphia PA, USA
>> International Centre for Theoretical Physics, Trieste. Italy.
>
>

-- 
Dr. Axel Kohlmeyer  akohlmey_at_gmail.com  http://goo.gl/1wk0
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.