*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
Dear Ian, exactly, I need to fit the smallest ellipsoid to the atoms, masses of which do not play any role. Thanks for pointing this out. That is why I was wondering whether someone has already solved that problem. Number of parameters: I would imagine that one has 3 parameters for the center of the ellipsoid (xyz), 3 parameters for the length of the three axes which are perpendicular to each other and another 3 parameters for the orientation of one of (the longest) axis i.e. 9 parameters total as you have stated, isn't this right? So, please, someone has already solved that? I'd appreciate any response. Marius > Marius - > > It depends on the precise problem you're trying to solve, i.e. do you > want just _any_ ellipsoid that encloses all the atoms, or do you > actually want the smallest ellipsoid (i.e. minimal volume). If the > former then the suggested method is sufficient, but note that there's > no > reason why the smallest ellipsoid should have the same centre or > principal axes as the inertial ellipsoid. > > If you want the precise minimal ellipsoid then it's a minimisation > problem. I've done it for the case of the smallest sphere enclosing > a > set of points, for the ellipsoid case it would be a matter of adding > 5 > more variables (i.e. 9 altogether) to the minimisation. Note that in > the sphere case as for the ellipsoid there's no reason why the > minimal > sphere should be centred at the centroid of the points (think about > the > case of a cluster of points plus 1 outlier). > > -- Ian > >> -----Original Message----- >> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On >> Behalf Of Tim Gruene >> Sent: 08 December 2005 16:10 >> To: Marius Schmidt >> Cc: [EMAIL PROTECTED] >> Subject: [ccp4bb]: Re: your mail >> >> *** For details on how to be removed from this list visit the *** >> *** CCP4 home page http://www.ccp4.ac.uk *** >> >> >> Hi, >> >> unless you need the ellipsoid oriented, you could use >> moleman2 to align the axes of inertia and then type 'stats' >> to see the extend of the molecule. >> >> Is this what you are looking for? >> >> Tim >> -- >> Tim Gruene >> Institut fuer anorganische Chemie >> Tammannstr. 4 >> D-37077 Goettingen >> >> GPG Key ID = A46BEE1A >> >> >> On Thu, 8 Dec 2005, Marius Schmidt wrote: >> >> > *** For details on how to be removed from this list visit the *** >> > *** CCP4 home page http://www.ccp4.ac.uk *** >> > >> > >> > Dear ccpbb'er, >> > is there anyone out there who has a chunk of code that can >> determine >> > an ellipsoid from a set of coordinates. >> > In this partiular example, a chromophore has to be >> inscribed into an >> > ellipsoid and the length of the 3 axes of the ellipsoid has to be >> > determined. Well, I could try to write the code myself, but maybe >> > someone encountered that problem before and can help me out. >> > >> > Looking forward to many responses >> > >> > M. Schmidt >> > >> > >> > >> > PD Dr. habil. Marius Schmidt >> > Physikdepartment E17 >> > Technische Universitaet Muenchen >> > James Franck Strasse >> > 85747 Garching/Germany >> > email: [EMAIL PROTECTED] >> > phone: +49-(0)89-2891-2550 >> > fax: +49-(0)89-2891-2548 >> > >> >> > PD Dr. habil. Marius Schmidt Physikdepartment E17 Technische Universitaet Muenchen James Franck Strasse 85747 Garching/Germany email: [EMAIL PROTECTED] phone: +49-(0)89-2891-2550 fax: +49-(0)89-2891-2548
