>> This is what RasMol/Chime use
>> (http://www.umass.edu/microbio/rasmol/rasbonds.htm).
>> But where did you get 0.45 A? R/C use 0.56 A according to the
>> above doc.

I read the source code, then read the rasbonds.htm page listed above.

The doc is consistent with the source code ... with one exception.

The 'slow' algorithm, which is applied when there are <= 255 atoms, uses a
maximum bonding distance which is:
  the sum of the two covalent radii + 0.56 angstroms.

The 'fast' algorithm, which is appled when there are >255 atoms, uses a
fixed maximum bonding distance of 475 rasmol units = 1.9 angstroms.

However, a different maximum bonding distance is used if either of the
atoms is a Hydrogen ... 300 rasmol units = 1.2 angstroms. This is what
keeps sibling hydrogens from getting connected, a concern I expressed in
my response to Eric's messsage. I did not see a reference to this special
treatment of hydrogens in the rasbonds.htm document (although I may have
missed it).

In both cases, the minimum is 100 rasmol units = 0.4 angstroms.


As to 'fast' vs 'slow' ... I find it hard to believe there was ever any
measurable performance difference between the 'fast' and the 'slow'
algorithms on any machine. In point of fact, the overall algorithm runs
*slower* because of the extra testing involved ... *if* we are using the
fast algorithm and *if* either of the atoms is a hydrogen.

I suspect that the 'fast' algorithm was implemented first, because it uses
fixed values and didn't need the covalent radius table. Later, the
covalent radius table was added. Rather than add 'set connect true', the
'fast' algorithm should have been eliminated at that point ... in my
perhaps-not-so-humble opinion :-)

Be that as it may, Jmol uses only one bonding algorithm, independent of
the number of atoms.


As to the 0.45 vs. the 0.56 ... the situation is a little more
complicated. The covalent radius values also enter into the equation, and
are probably more important than the 'tolerance' factor. We couldn't
change the tolerance factor to 0.56 value without also adopting the RasMol
covalent radius values.

Frankly, I feel more comfortable with the OpenBabel covalent radius values
and tolerance factor. They are more modern, calculating bonds properly is
part of the OpenBabel charter, and it seems to me that they should have a
broader base of experience which extends beyond biomolecules.

I would like to stick with the OpenBabel values unless/until someone comes
up with a 'problem case' ... at which point we should tell the OpenBabel
team to fix it.


Miguel





-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
_______________________________________________
Jmol-developers mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jmol-developers

Reply via email to