[gmx-users] fixed number of clusters from a MD simulation

[Tsjerk Wassenaar]

Hi Rebeca,

You'd need to use k-means/k-medians clustering for that, which is not in
gmx cluster. I'd suggest to perform PCA and collect the projections onto
the first ten or so eigenvectors to decrease the dimensionality. This data
matrix you can then process with, e.g., R to perform k-means clustering.

Cheers,

Tsjerk
On Mar 31, 2015 2:00 AM, "Rebeca García Fandiño" <regafan at hotmail.com>
wrote:

> Dear Gromacs users,
> I am trying to find a method to obtain a fixed
> number of clusters (using g_cluster) from a MD simulation. The search of
>  clusters is very "manual", depending a lot of the cutoff selected for
> the analysis. Is there any way to calculate a fixed number of clusters
> from a MD simulation, in such way that the program calculates the proper
>  cutoff to provide the exact number asked by the user?
> Best wishes,
> Rebeca.
>
> Dr. Rebeca Garcia
> Santiago de Compostela University
> Spain
>
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>