[gmx-users] g_clustsize and monomers

[devicerandom]

On 27/03/11 20:55, David van der Spoel wrote:
> On 2011-03-27 21.27, devicerandom wrote:
>> On 27/03/11 16:29, David van der Spoel wrote:
>>> On 2011-03-27 16.42, devicerandom wrote:
>>>> Hi,
>>>>
>>>> I am trying to use g_clustsize to follow aggregation of peptides. I am
>>>> currently using the following line:
>>>>
>>>> g_clustsize -f my.trr -s ../my.tpr -nc nclusters.xvg -mol
>>>>
>>>> Problem is, it actually counts monomers as clusters. I have a system
>>>> with 125 separate molecules, and the number of clusters starts from
>>>> 125.
>>>>
>>>> What can I do to indicate a minimum cluster size?
>>>>
>>>> m.
>>>>
>>> What does it matter? Just simulate longer until you get two peptides to
>>> aggregate. Check eg. maxclust.xvg.
>>>
>>
>> Apologies, I've probably not been clear (I was writing in a rush, my
>> fault).
>>
>> It matters because they aggregate nicely, but I want to follow the
>> number of *clusters* (that is, assemblies of 2 or more molecules) in
>> time (More precisely, I want to know the number of molecules within
>> aggregates).
>>
>> Any hint? g_clustsize is a bit undocumented apart from the help page.
>>
> try the xpm files, and process those with xpm2ps

Hm. They seem to be kind of very raw histograms of cluster size 
distribution. Not exactly what I'm looking for.

What I'd like is something exact, like histo-clust.xvg, but for each 
frame (while histo-clust.xvg seems just to provide some kind of, er, 
average histogram on all the trajectory? counts are not integer)

I understand there is nothing that does it apparently, but at least 
having some output to script that from would be better than tinkering 
directly with g_clustsize code (my C skills are not exactly stellar).

thanks!
m.

-- 
http://devicerandom.org