Re: [AMBER] clustering in PC subspace

From: Daniel Roe <>
Date: Fri, 30 Jan 2015 14:24:59 -0700


You could perform clustering with PC projections using cpptraj. The
following is a cpptraj script that I have used to cluster via PC
projections. You'll have to adjust masks, # of vecs, epsilon, etc for
your system of course. Be careful when using the 'loadpairdist'
keyword since any changes to your distance metric (including sieve
value) will invalidate the saved pairwise matrix; cpptraj catches most
of these cases but may not catch all.

# Average the trajectory.
rms first :1-12.CA
average Avg.rst7 restart
createcrd crd1
# Fit to average structure, save fit coordinates.
reference Avg.rst7.1 [avg]
crdaction crd1 rms ref [avg] :1-12.CA
# Calculate eigenvalues/eigenvectors.
crdaction crd1 matrix covar :1-12.CA name myCovar
runanalysis diagmatrix myCovar out evecs.dat vecs 36 name MyModes
# Project fit coordinates onto eigenvectors.
crdaction crd1 projection P1 modes MyModes beg 1 end 36 :1-12.CA
# Perform clustering using projections.
runanalysis cluster C0 data P1:1-36 crdset crd1 \
        hieragglo epsilon 11.5 averagelinkage \
        sieve 10 out cnumvtime.dat repout rep repfmt pdb \
        summary summary.dat info info.dat cpopvtime cpop.agr normframe \
        loadpairdist pairdist CpptrajPairDist \
        singlerepout singlerepfmt netcdf

Hope this helps,


On Fri, Jan 30, 2015 at 4:09 AM, Asmita Gupta <> wrote:
> Dear users,
> Is there any way to perform clustering in PC subspace in AMBER?...i know of
> a program Bio3D that can do that, but can that be done with AMBER?
> Thanks
> Asmita
> _______________________________________________
> AMBER mailing list

Daniel R. Roe, PhD
Department of Medicinal Chemistry
University of Utah
30 South 2000 East, Room 307
Salt Lake City, UT 84112-5820
(801) 587-9652
(801) 585-6208 (Fax)
AMBER mailing list
Received on Fri Jan 30 2015 - 13:30:03 PST
Custom Search