The PROTEIN program system
W. Steigemann
- Rechenzentrum
- Max-Planck-Institut für Biochemie
- D-82152 Martinsried
- Germany
- Tel: +49 (89) 8578-2723
- Fax: +49 (89) 8578-2479
- e-mail: steigema@biochem.mpg.de
The PROTEIN program system is an integrated collection of
crystallographic programs designed for the structure analysis of
macromolecules. In addition to our own programs, some routines kindly
supplied by various authors have been implemented into a common
environment to ensure a homogeneous and user-friendly treatment of
command input, file handling and memory management. Although the
fastidious name PROTEIN has come into use within our institute for the
program system, it does not claim to completeness. There are still a
number of stand-alone programs in use which are compatible with the
system only on the file-handling level.
The most important tasks of the program are:
- Generation and expansion of data files with reflection data
- scaling of reflection data of different crystals or films onto a
common scale
- averaging the reflection data and elimination of inaccurate or
obviously wrong measurements with calculation of R-values describing
the quality of the measured data
- scaling of derivative data
- calculation of Patterson, difference Patterson, Fourier, difference
Fourier maps with many options by normal and fast Fourier
transform algorithm (method of L.F. Ten Eyck)
- M.I.R. (multiple isomorphous replacement) and heavy atom parameter
refinement (based on a program originating from M.G. Rossman,
extended by L.F. Ten Eyck and S.J. Remington)
- listing, contouring (also stereo), peak searching of 3D maps in all
directions of the crystal axes
- fast calculation of structure factors from atomic coordinates
- statistical supplement, e.g. calculation of distribution of figure
of merit, significance of anomalous dispersion data, crystallographic
R-value, etc.
- real space search methods, e.g. self- and cross rotation,
translation function using Patterson and Fourier maps, rotation of
Fourier maps, vector verification as aid in the interpretation of
difference Patterson maps, etc.
Much care has been taken to keep the programs in a general form in
order to treat all user specified crystallographic space-groups.
The PROTEIN program intentionally does not contain links for:
- Refinement. A number of methods are available (e.g. XPLOR, TNT,
...).
- Interpretation of structural details from stereo chemical
considerations. For this purpose interactive graphics systems (e.g.
FRODO, O) are most useful.
The program system over its almost 20 years of existence (Ph.D. Thesis
W. Steigemann) has become wide spread within the crystallographic
community (some 150 installations worldwide). In fall 1991 (initiated
during a short-term stay in J. Deisenhofer's laboratory at the HHMI in
Dallas in summer 1989) a major release (V3.1) of the package has been
launched with an enhanced file structure (provision for standard
deviations of observations, several phase sets, anisotropic scaling
etc.) and the additional function of previously separate programs. These
include:
- use of standard deviations (selection of significant reflections,
direct use in scaling, averaging, M.I.R. phasing)
- classification of reflections
- addition of partially recorded reflections
- elimination of outliers of derivative data on the basis of
comparison with native data
- phase encoding of various phase information in Hendrickson-Lattman
coefficients
- phase combination at various instances (e.g. M.I.R. phasing, export
of data, statistical functions)
- very general program for setup of Fourier coefficients with the
ability of phase combination
- direct support of PROTEIN like map files from real space search
routines
- enhancements in statistical programs (phase reliability and
distribution, absolute scaling etc.)
- indirect command files in the input command stream
Major functional enhancements are planned for the next release V3.2
(partly available already). These include, in particular,
- electron density modification techniques (e.g. solvent flattening,
non-crystallographic symmetry averaging)
- use of observed e.s.d.'s for the scaling and averaging of intensity
data
- enhanced statistics for the scaling and averaging of observed
intensity data
- contouring of rotation function maps
- simple space group alterations in existing primary data files
- further extensions in the exchange of structure factor and map data
with "foreign" programs
- generation of Postscript graphical output
- enhanced documentation
An important issue particularly addressed in the new version has been
the portability to different hardware platforms. Following the current
trends in computer technology with the availability of high performance
cost-effective RISC workstations the program has been ported to Unix.
The program is currently distributed for operation on computers and
workstations of the manufacturers DEC (VAX/VMS, OpenVMS, OSF/1), SUN,
SGI, E&S and CONVEX (Cxxx).