CCL: clustering docked poses

There is some ambiguity about which set is to be translated and rotated.
I did not mean that the ligand is moved and the enzyme is fixed.
If so, a rigid ligand would be perfectly superposed to any translated
and rotated copy of itself.
I meant that the whole complex ligand-enzyme, when compared  to
an other complex of the same ligand and the same enzyme,
needs that the RMSD is computed in a common set of cartesian
coordinates. This is not ensured, in general.
So, either the enzyme is superposed to itself (the minimized RMSD
is not null if the enzyme structure changes, even slightly), or,
the full complexes are superposed (the minimized RMSD is not null
unless the complexes are identical).
About the pairwise correspondence: this latter is read as an input data
by ARMS, but it is computed as an output result by CSR (returns also
the maximal common 3D motif).
Other debate: I do not encourage the use of Excel, Word, etc.
which use non standard binary formats not suitable for professional
uses such as reading files by our own programmes (say, in f77, or in C,..),
converting formats, exporting data to other applications, etc.

All the best,

Michel Petitjean,
DSV/iBiTec-S/SB2SM (CNRS URA 2096), CEA Saclay, bat. 528,
91191 Gif-sur-Yvette Cedex, France.
Phone: +331 6908 4006 / Fax: +331 6908 4007
E-mail: michel.petitjean]-[cea.fr, petitjean.chiral]-[gmail.com
http://petitjeanmichel.free.fr/itoweb.petitjean.shape.html
.......................................................................................................
 Sent to CCL by: maxim totrov [max*o*molsoft.com]
For docking poses (as opposed to conformational analysis), you usually do not want to minimize RMSD for rotations and translations (RMSD with superposition). As an extreme case, consider completely rigid ligand: you still might have many distinct docked poses, but RMSD with superposition will always be zero.
On the other hand, an important issue arises if there are symmetric moieties in your molecules - one has to consider different possible correspondences between equivalent atoms in two poses. Think of a flipping phenyl moiety for example. Excel will not detect that. Among other software, Molsoft's free Browser can calculate either static or superimposed RMSDs, with or without consideration of symmetries. Scripting language allows easy batch processing. 
 Best regards,
 Max Totrov

 
 Sent to CCL by:
 "PETITJEAN Michel" [michel.petitjean:cea.fr]

  Does Excel minimizes the RMSD for all rotations and translations ?
 If not,
 you need a software such as ARMS or CSR.
 Michel Petitjean

 DSV/iBiTec-S/SB2SM (CNRS URA 2096)
 CEA Saclay, bat. 528
 91191
 Gif-sur-Yvette Cedex, France.

  Phone: +331 6908 4006 / Fax: +331 6908 4007
 E-mail: michel.petitjean^^cea.fr
 http://petitjeanmichel.free.fr/itoweb.petitjean.freeware.html