Re: software/algorithm to cluster protein conformations

From: Nicholas M Glykos (glykos_at_mbg.duth.gr)
Date: Fri Apr 30 2010 - 04:40:02 CDT

> I would like to cluster different conformations of a protein focusing on
> certain residues and their side chains. Is there any available
> softwares/algorithms that would perform this?

If you assign to the atoms you are interested-in a distinct SEGID,
say L, then you can use

# carma -v -w -col -cov -eigen -dpca 5 3 298 -segid L protein.dcd protein.psf

for performing a PCA-based clustering in dihedral space (which ignores
side chains), or better,

# carma -v -w -col -cov -eigen -proj 5 3 298 -segid L -atmid HEAVY protein.dcd protein.psf

for performing a PCA-based clustering (using all heavy atoms) in Cartesian
space. If you'd rather be safe than sorry, you can use something like

# crossDCD protein.psf protein.dcd protein.dcd 200 " -segid L -atmid HEAVY "

to prepare a square matrix with all frame-to-frame rmsds between the
selected atoms (all heavy atoms of SEGID L) which you can then use with a
statistics program (like R) for cluster analysis. Both carma and
the crossDCD script are available from http://utopia.duth.gr/~glykos/Carma.html

Nicholas

-- 
          Dr Nicholas M. Glykos, Department of Molecular Biology
     and Genetics, Democritus University of Thrace, University Campus,
  Dragana, 68100 Alexandroupolis, Greece, Tel/Fax (office) +302551030620,
    Ext.77620, Tel (lab) +302551030615, http://utopia.duth.gr/~glykos/

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:22:56 CST