#585 The LengthOverBreadth incorrectly gets VdW radii

cdk-1.2.x
closed
nobody
9
2012-11-03
2008-01-02
No

It seems to have sort of depended on atom type names to be unset, because if set, e.g. using the CDKAtomType perception, the factory will not find any atom type in the Jmol list... The descriptor did not do any perception itself...

I have updated this code to pick the first Jmol atom type matching the atom's symbol, which I think was the old behavior...

However, instead, it should do proper atom type perception...

Discussion

  • Rajarshi Guha

    Rajarshi Guha - 2008-01-02

    Logged In: YES
    user_id=349408
    Originator: NO

    What do you mean by proper atom type perception?

    Does this imply that for every atom typing scheme we need to have a set of van der waals radii? I would think not since VdW radii should not depend on atom typing (since they're supposed to derived from a hard sphere model of atoms in an unbound state). As a result, VdW should be something like a look up table based on symbol (may be part of the Periodic Table?)

     
  • Rajarshi Guha

    Rajarshi Guha - 2008-06-11

    Logged In: YES
    user_id=349408
    Originator: NO

    Was there any updates on this? I still favor the idea of moving VdW radii to something that is not dependent on atom types

     
  • Egon Willighagen

    Logged In: YES
    user_id=25678
    Originator: YES

    No, the code has not been fixed yet. But VdW is now separately, so this code can be 'fixed'.

     
  • Rajarshi Guha

    Rajarshi Guha - 2008-07-15

    Logged In: YES
    user_id=349408
    Originator: NO

    However, the code actually uses covalent radii, rather than VdW radii. Since covalent radii will depend on atom type, perception is required.

    However, my opinion is that using VdW radii should be fine - this is not a high resolution descriptor and therefore, using fixed VdW radii (i.e., ignoring atom types) is probably not a big issue.

     
  • Egon Willighagen

    Rajarshi, what's the status of this bug?

     
  • Rajarshi Guha

    Rajarshi Guha - 2008-10-19

    It looks like we should simply include reported covalent radii in the periodic table class, rather than have it depend on atom type. We should probably just use the BODR data (or maybe a recent source such as
    http://dx.doi.org/10.1039/b801115j )

     
  • Rajarshi Guha

    Rajarshi Guha - 2008-10-20

    Updated to use the BODR covalent radii , no more dependence on atom type based covalent radii

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.





No, thanks