[Molvis-list] exporting _good_ h-bonds (was: export Chime h-bonds)

Roger Sayle roger at eyesopen.com
Tue Jan 3 11:06:42 EST 2006

On Tue, 3 Jan 2006, Kevin Karplus wrote:
> Although RASMOL *claims* to use the DSSP algorithm for identifying
> Hbonds, it does not seem to use the entire algorithm as it reports
> many erroneous H-bonds that DSSP omits on geometric grounds.
> Rasmol seems to use only the distance and not the geometry of the hbond.

I'm pretty certain that RasMol's implementation is an accurate
implementation of the original DSSP algorithm.  In the Kabsch
and Sander definition, the existence of a structural hydrogen
bond is not based directly on either distance or geometry, but
as a threshold on the coulombic interaction energy.  It is the
relative interaction energies of H..O, N..O, H..C and N..C that
defines the acceptable geometries.

The interested reader can refer to the distance vs. angle graphs
provided in Fig. 1 of "Dictionary of Protein Secondary Structure:
Pattern Recognition of Hydrogen-Bonded and Geometrical Features",
Wolfgang Kabsch and Christian Sander, Biopolymers, Vol. 22,
pp. 2577-2637, 1983.

Admittedly, the Stride definition of a backbone hydrogen bond
given in "Knowledge-based Protein Secondary Structure Assignment",
Dmitrij Frishman and Patrick Argos, Proteins: Structure, Function,
Genetics, Vol. 25, pp. 566-579, 1995 is more advanced, but at
the time RasMol was written DSSP definitions were clearly the
de facto standard.

If anyone knows of a discrepancy between between RasMol's
secondary structure assignment and those assigned by DSSP,
let me know and I'll investigate.  Be warned that in the
Kabsch and Sander definition there may be numerous backbone
hydrogen bonds that are then not recognized as forming part
of a standard secondary structure unit.

For protein sidechain and protein-ligand hydrogen bonding, one
might consider a more general definition such as the one used
in GRID: D.Boobbyer, P.Goodford, P.McWhinnie and R.Wade,
"New Hydrogen-Bond Potentials for Use in Determining Energetically
Favorable Binding Sites on Molecules of Known Structure",
Journal of Medicinal Chemistry, Vol.32, No.5, pp.1083-1094,

Roger Sayle,                         E-mail: roger at eyesopen.com
OpenEye Scientific Software,         WWW: http://www.eyesopen.com/
Suite 1107, 3600 Cerrillos Road,     Tel: (+1) 505-473-7385
Santa Fe, New Mexico, 87507.         Fax: (+1) 505-473-0833

More information about the Molvis-list mailing list