Ksdssp is an implementation of the Kabsch and Sander algorithm for defining the secondary structure of proteins, as described in:
W. Kabsch and C. Sander, "Dictionary of Protein Secondary Structure: Pattern Recognition of Hydrogen-Bonded and Geometrical Features" Biopolymers 22:2577 (1983).Many, but not all, input structure files include protein secondary structure information. In PDB format, secondary structure assignments are described in HELIX and SHEET records (see the Intro to PDB Format (PDF)). When a peptide or protein structure file that does not include secondary structure assignments is read, ksdssp is automatically invoked to generate helix and sheet information. Ksdssp is also called automatically when MatchMaker is used with the Compute secondary structure assignments option, regardless of whether secondary structure assignments already exist.
In Chimera, helix and sheet assignments are stored as true/false settings (for each amino acid residue) of the residue attributes isHelix and isStrand (isSheet). The assignments are used for drawing protein ribbons and may be used by MatchMaker, depending on its settings.
Parameter settings used are reported in the Reply Log. Defaults to be used when ksdssp is called automatically can be changed with the compute SS dialog (opened from the Model Panel). Defaults for the ksdssp command, however, cannot be changed and are given below. Different values can be specified with the command-line options.
Ksdssp uses the coordinates of the backbone atoms (N, CA, C, O, and optionally H) of a protein to determine which residues are in alpha helices and beta strands. If an amide hydrogen is missing, it is placed 1.01 angstroms from N along the bisector of (1) the vector opposite the bisector of C-N-CA, and (2) the vector opposite the C-O vector from the previous amino acid.
The calculation is applied to the models containing atom-spec, and a blank atom-spec is interpreted as "all." Models that do not appear to contain amino acids (that is, nonpeptide molecules and nonmolecular objects) are ignored.
Ksdssp estimates the energy of each candidate hydrogen-bonding interaction and classifies it as a hydrogen bond if the energy is no greater than (at least as favorable as) cutoff. The default is –0.5 kcal/mol, as recommended by Kabsch and Sander, who add that "A good H-bond has about -3 kcal/mol binding energy."
HELIX records are only generated for alpha helices at least helix_min residues long; the default is 3.
SHEET records are only generated for beta strands at least strand_min residues long; the default is 3. Reducing strand_min to 1 is not recommended, as there are bridges in many structures that confuse the algorithm for defining sheets.
Save the hydrogen-bonding information in file rather than discarding it (not yet implemented in Chimera). The notation in the file is similar to that used by Kabsch and Sander, but is in a vertical instead of horizontal format.
In most cases, the default parameter values are reasonable for computing secondary structure.
See also: ribbon, PipesAndPlanks, MatchMaker