[gmx-users] cluster analysis

[UCT Staff Member - Jackson]

Dear All

I have done a high temperature dynamic search of conformational space of
a cyclic peptide collecting 1000 structures on the way.  I then used
trjcon -sep to split the trajectory into individual conformations which
I then annealed.  I now want to do cluster analysis of these annealed
structures.  They all have the same simulation time so I used trjconv to
reintroduce the time they were collected in the dynamics run.  I don't
believe this has any meaning but it differentiates the different
structures during the cluster analysis.
After the cluster analysis I get 1 cluster and the time of the middle
conformation.  However, I want the lowest energy conformation in this
cluster.  Is there any way of getting this or of listing the energy of
each of the conformations?  Also is there a way of differentiating the
conformations by frame number rather than time?
I want to do several dynamic runs with different starting conformations
and the pool them all together for a single cluster analysis.

Thanks for the help

Graham
-- 
Professor Graham E. Jackson
Department of Chemistry
University of Cape Town

Tel: (021) 6502531
Fax: (021) 6897499