superimposing molecules

David Jones jones at
Wed Jun 5 12:16:10 EST 1996

Andrew Dalke (dalke at wrote:

: I thought the standard way to do rmsd best-fit alignment of two structures
: with 1-1 assignment, as in Kabsch, Acta Cryst. (1978) A34, 827-828,
: and that's what I used in our visualization program, VMD.  It isn't
: iterative and works fine as far as I can see.

Well, it's a commonly used method, granted, but I don't think there is
a "standard" way to superpose two structures. It's not much different
from the variety of methods for curve-fitting, matrix diagonalization, or
just about any numerical computing problem.

I must have seen at least 10 papers (and there must be more) in Acta Cryst
detailing different ways of skinning this particular cat - Andrew McLachlan
has published two alternatives, for example - one based on U matrix
diagonalization (1979) and one based on conjugate gradient minimization
(1982). Just doing a quick scan through the program codes on my
machine here, I can see at least 5 different approaches used by
different authors. I guess some must be quicker than others, some more
stable as the equivalent points approach colinearity, but they all seem
to work OK in the applications concerned.

This message was written, produced and executively directed by Dr David Jones
New Address: Dept. of Biological   |   Email: jones at
Sciences, University of Warwick,   |     Tel: +44 1203 523729
Coventry CV4 7AL, U.K.             |     Fax: +44 1203 523568

More information about the Molmodel mailing list