I believe that what you are looking for is called singular value
decomposition. See
http://www.rrz.uni-koeln.de/REDUCE/linalg/node57.html
- Jonathan
Jonathan Epstein epstein at ncbi.nlm.nih.gov
National Center for Biotechnology Information Phone: (301)496-2477 x254
National Library of Medicine Building 38A, Room 8N805
National Institutes of Health 8600 Rockville Pike
Bethesda, MD 20894
islam (islam at icrf.icnet.uk) wrote:
> I would be grateful for any ideas, comments or software to deal
> with the following problem please:
> Given an orthogonal dataset of coordinates for 2 identical
> molecules is it possible to derive the symmetry operator relating
> them or determine if one exists ?
> Am I correct in thinking:
> * the problem is equivalent to least squares fitting the
> two molecules to derive a transformation matrix and then
> decomposing/examining the derived matrix to its "simplest
> form" e.g. rotation about a 3D axis
> * in a sense as long as a transformation matrix can be obtained
> to fit one molecule exactly onto another, then a symmetry op
> always exists defined by the matrix ?
> Can the above idea be extended to derive the crystallographic
> symmetry operator relating the two molecules (obviously given
> the appropriate cell,space group etc). (An alternative here,
> I guess, would be to generate unit cells to find which symmetry
> elements related to the 2 molecules).
> Thanks.
> ____________________________________________
> Suhail A Islam
> Biomolecular Modelling Laboratory
> Imperial Cancer Research Fund, P.O. Box 123
> 44 Lincoln's Inn Fields, London WC2A 3PX
> Tel: (0171) 269 3380, Fax: (0171) 269 3479
> email: islam at icrf.icnet.uk>http://www.icnet.uk/bmm/> ____________________________________________
