ADCGAL
cutoff = maximum limit on perturbation (in the same units as the coordinates, i.e.
Angstrom for proteins). This is the parameter epsilon in my definition
of AD(epsilon). To try it out, a typical default value is 1.0 Angstrom,
though values from 0.01 to 2.0 Angstrom are useful for capturing different
ranges of perturbation in proteins.
prune = maximum edge length allowed in a simplex. This gets rid of long edges on the
convex hull and within interior pockets and cavities. A typical default
value in proteins is 10 (Angstrom), though values from 6.0 to 15.0 Angstrom
may be useful. "SHORT" edges, triangles and tetrahedra are those with all
edges below the prune in length.
file prefix = 1xyz if your 3D coordinates are stored as rows of three coordinates
in a file "1xyz.out". This parameter can include a full path if the
.out file lies elsewhere. If this parameter is missing, input comes
from the standard input, and te prefix "points" is made up for purposes
of writing output (hence you can use AD as an input pipe, though
minor changes are required to use it as an output pipe).
Output: If the file prefix for input was "1xyz", output is written to files as follows:
1xyz.del: short Delaunay edges, written as rows of SORTED point index
pairs starting at 0; example row "0 4"
1xyz.AD: short AD edges with threshold>0, written as rows of sorted edge
index pairs and a threshold. Example row: 5 6 0.7773
The files below are generated only by the version of ADCGAL that calculates
simplices, and not by the edge version (the symbol ADSIMP_CGAL has to be
defined during compilation.
1xyz.del3: Delaunay triangles with short edges, rows of 3 sorted indices
1xyz.AD3: AD triangles, rows of 3 sorted indices and a threshold
1xyz.del4: Delaunay tetrahedra, 4 indices
1xyz.AD4: AD tetrahedra, 4 indices and a threshold
Visualizing AD edges, triangles and tetrahedra:
Please refer to the README.ADMATLAB.txt, where there is some spiffy code (plotAD.m, plotAD4.m,
plotADspheres.m, and the kinemage converters) to visualize these AD tetrahedra, triangles or
edges on top of the set of points. It is not too hard to write your own visualizer also,
in MATLAB, C++ or anything else, remembering to index the points from 1xyz.out and treating
the indices in *.del*,*.AD* files as a C-style index starting at 0 (hence adding 1 if you are
in MATLAB).