|
590. DGEOM(95): Distance Geomerty Pgm
Copyright 1990 DuPont Corporation
Copyright 1995 Chiron Corporation
All Rights Reserved
**This software is proprietary and confidential to Chiron
Corporation**
**and is furnished under a license agreement. This
software may be**
**used and disclosed only in accordance with the terms of
such license.**
Authors: Jeffrey M. Blaney, Gordon M. Crippen, Andrew
Dearing, J. Scott Dixon,
and David C. Spellmeyer, 1984-1994.
Automatic double bond assignment is based on routines
provided by Eric
Swanson, University of Washington. The random number
generator originally
appeared in "Toward a Universal Random Number Generator" by
George Marsaglia
and Arif Zaman, Florida State University Report: FSU-SCRI-
87-50 (1987). It
was later modified by F. James to produce an array of
pseudorandom numbers.
The conjugate gradient minimizer, CONMIN, comes from D. F.
Shanno and K. H.
Phua, ACM Transactions on Mathematical Software, Vol. 6
(December 1980),
618-622.
Description
DGEOM is a distance geometry program for molecular model-
building, receptor
modeling, conformational analysis, and NMR solution
structure determination,
using the algorithms developed by Crippen, Havel and
Kuntz^2-5. Distance
geometry is a general method for converting a set of (NxN)-N
distance bounds
into a set of 3xN Cartesian coordinates consistent with
these bounds.
Molecular structures can be described by the set of all
pairs of interatomic
distances. For example, the distance matrix for cyclohexane
is:
1 2 3 4 5 6
1 0.0 1.5 2.5 3.8 2.5 1.5
2 1.5 0.0 1.5 2.5 3.8 2.5
3 2.5 1.5 0.0 1.5 2.5 3.8
4 2.6 2.5 1.5 0.0 1.5 2.5
5 2.5 2.6 2.5 1.5 0.0 1.5
6 1.5 2.5 2.6 2.5 1.5 0.0
The upper diagonal contains the upper bounds (maximum
distance) and the lower
diagonal contains the lower bounds (minimum distance)
allowed between each
pair of atoms. For this example, the upper and lower bounds
are equal except
for 1-4, 2-5 and 3-6, since all other pairs involve atoms
which are covalently
bonded to each other or form a bond angle (bond lengths and
angles are fixed).
The lower and upper bounds for 1-4, 2-5 and 3-6 are set to
values correspon-
ding to a cis or trans conformation, respectively.
This is the first update of the original DGEOM system (QCPE
590) since 1990.
This is a fairly general UNIX version but should readily
port to almost any
modern computer. The following is a list of improvements
and bug fixes which
have accumulated since 1990:
* Partial metrization algorithm improves random sampling
(J. Kuszewski, M. Nilges, A. T. Brunger, "Sampling and
Efficiency of Metric Matrix Distance Geometry: A
Novel
Partial Metrization Algorithm", J. Biomol. NMR, 2, 33-
56
(1992)).
* New lower-bound error function improves convergence
(T.
F. Havel, "An Evaluation of Computational Strategies
for
Use in the Determination of Protein Structure from
Distance
Constraints Obtained by Nuclear Magnetic Resonance",
Prog.
Biophys. Molec. Biol., 56, 43-78 (1991)).
* Molecular dynamics has been added as a refinement
method to supplement conjugate gradient minimization.
* Random coordinate generation has been added as an
alternative to embedding.
* Automatic connectivity generation and double-bond
assignment (even for conjugated and aromatic systems).
This eliminates the need for supplying CONECT records
and/or editing them to specify double bonds for most
organic structures, including peptides and
nucleotides.
Automatic connectivity generation will be performed if
CONECT records are missing.
* Free-format input for CONSTRAINT and ORIENT files.
Backwards compatible with original version, except
that now
a "*" is required for molecule and residue wildcards
to
mark their fields (previously, a blank was used).
Original
DGEOM 1990 fixed-format constraint files will still
work.
* Free-format input of van der Waals and covalent bond
radii (the latter are needed for connectivity
generation).
DGEOM preassigns van der Waals and covalent bond radii
only
for C, N, O, S, P, H, F, Cl, Br, I and Du (dummy atom,
radius = 0).
* REJECT rejects similar structures. REJECT rejects the
current structure if its minimum RMSD to all
previously
accepted structures is less than REJECT's threshold
(default = 1.0 angstrom). This ensures that all
output
structures will have a minimum pairwise RMSD >
REJECT's
threshold.
* New rules to detect conjugated groups (e.g.,
butadiene,
aniline or phenol-like systems which have restricted
rotation about the single bond) and constrain them to
be
nearly coplanar (+/- 40 degrees); disulfide and
hydrazine
torsions are constrained to be +/- 90 degrees.
Recognizes
thioamides and thioesters.
* Read atom names beginning with integers and residue
numbers with optional chain identifiers from PDB
format
files.
* New FLOAT command available in CONSTRAINTS file for
specifying chiral centers that can racemize or
"float", and
amide or ester carbonyl carbons that should randomly
sample
cis and trans.
* Warns about newly assigned, weaker constraints that
overwrite previously assigned, stronger constraints.
* Molecule-molecule, molecule-residue or residue-residue
distance constraints are now all available.
* Wildcard selection of all atoms in a molecule with
GROUP, MATCH or PLANAR constraints is now possible.
* PDBOUT can now specify the starting output structure
number by appending a ":" to the root filename,
followed by
the desired number.
* PREFIX toggles writing the structure number before or
after the fileroot when writing out PDB files.
* SPLIT writes each molecule to a separate file with
as
expected with "PREFIX:".
* Modified all Boolean commands to take TRUE/FALSE
option; entering a command with no value implies TRUE.
* Comment lines can be included in DGEOM command files;
summary
output file.
* Chiral constraints can be reinitialized, clearing all
previously set chiral constraints (automatic and user-
defined) with the new command: CHIRAL INIT.
* User-defined random number generator initialization.
NSEED1 and NSEED2 set the 1st and 2nd seeds for the
random
number generator, and NRANDOM initializes the random
number
sequence to the NRANDOMth element.
* ORIENT command warns if fewer than 4 atoms are entered
for matching.
* Program COMPARE includes Jarvis-Patrick1 clustering
and
is now completely self-contained. COMPARE reads a
series
of PDB files generated by DGEOM, along with (optional)
a
free-format file containing the residue number and
names of
atoms to match. It calculates all pairwise RMS least-
squares fit errors between the structures, generates a
distance matrix from the RMS matrix, extracts a set of
nearest neighbors from the distance matrix, and
clusters
the nearest neighbor list using the Jarvis-Patrick
algorithm to group the structures into conformational
"families".
Bug Fixes
* Fixed bug to allow read/write of 4-character residue
names.
* Fixed bug which forced the end of a constraints group
after reading a COMMENT card or blank line.
* Fixed bug in the message warning of more than 4 bonds
to a given atom.
* Fixed bug involving read/write of the metric matrix to
a scratch file.
* Fixed bug where automatic chiral constraint generation
could overflow the format statement in an internally
generated chiral command.
* Chiral quartets within RIGID groups are now only set
if
the abs (chiral volume) threshold, where threshold ~
max
volume for a chiral carbon. Previously, very large
volumes
between atoms far apart in a structure could lead to
severe
convergence problems.
* Fixed bug in eigenvalue routine for two-dimensional
structures.
* Fixed bug in torsion picking routine where upper and
lower bounds are nearly equal.
* Fixed bug in routine to find aromatic rings.
* Fixed sulfate/phosphate bug where planar constraint is
mistakenly assigned to these groups.
| 
