Top page

About the MolScript program

Publication reference

The MolScript program is described in the following publication:

Per J. Kraulis
MOLSCRIPT: A Program to Produce Both Detailed and Schematic Plots of Protein Structures.
Journal of Applied Crystallography (1991) vol 24, pp 946-950.

This citation must be given in all publications using images produced with MolScript. The paper is available at the IUCr Journals website.

Notes on some features

Hermite splines

The spline function used to create segment coordinates for various objects (coil, turn, strand, helix) is the Hermite spline (Foley et al 1990). This is based on two control points, which the curve passes through, and directional vectors at these points. The advantage of this spline is that it passes through all its control points. However, it requires direction vectors of defined lengths. In MolScript, such vectors are computed in different ways depending on the graphical object.

Irregular secondary structure

A feature of MolScript is that irregularities in helices and strands are still visible to some degree in the schematic drawings. Especially the helices manage to convey bends and other irregularities without turning ugly. Since only CA-coordinates are used for creating the schematic drawings, even poorly refined structures can give reasonable-looking schematic objects. This may or may not be advantage, depending on one's point of view...

Selection mechanism

The atom and residue selection mechanism was inspired by a similar facility in the structure refinement program X-PLOR (Brünger 1992).

Orthographic projection

An orthographic projection is used in all output modes except VRML 2.0 output mode. The projection used in VRML cannot be specified by MolScript; it is entirely dependent on the VRML browser used to display the VRML file.

Background

The MolScript program was conceived in the tradition of molecular graphics of Jane Richardson, Arthur Lesk & Karl Hardman, and John Priestle. The aim was a program that could produce nice schematic drawings of proteins as well as detailed views of specific residues. The graphical objects helix, strand and coil were designed to look as much like Jane Richardson's drawings as possible.

The success of the MolScript program (more than 2,000 academic licenses, and more than 30 commercial licenses issued by October 1997) shows that there was a great need for such a program. In the years since MolScript was first released (1991), several other similar programs have appeared, which either are designed solely for producing high-quality images of biological macromolecules, or have such output procedures as integral features. For MolScript to remain useful in such competition, novel features and improvements in the software in the light of experience had to be implemented.

The emergence of the VRML 2.0 standard in 1996-1997 for 3D Web applications was an inspiration for creating a new version of MolScript, since it was obvious that the features in MolScript were a very good basis for creating informative and useful VRML representations of molecular structures. The Web has from its start been employed by the molecular biologists and geneticists for sharing data and knowledge, thereby helping to create the new field called bioinformatics. There is a huge potential for the biological structural sciences to use the Web for sharing not only the structural data itself, but also structural knowledge. The VRML 2.0 standard is a good basis for this.

The appearance of the OpenGL standard library and API for high-quality 3D graphics applications in 1993 made it possible to write an interactive-graphics implementation for MolScript which potentially could run on a broad range of computer systems. Also, the off-screen rendering capabilities of many OpenGL implementations made it possible to produce image files directly through MolScript. Since such image files are now a staple of the Web, such a feature would clearly be quite useful.

Implementation

The internal program structure of the previous version of MolScript (v1.4) did not make it easy to introduce new output formats. When an attempt was made to create a VRML 2.0 interface in a modified version of MolScript, it became clear that a redesign was necessary.

The previous versions of MolScript were written in Fortran 77. One major problem with this was the lack of a standard dynamic-memory feature in Fortran 77. This and other problems were reasons for choosing another programming language.

Considering the various available programming languages, the final decision was to use ANSI C: it is a small, simple language, with almost universal and standardized support on all relevant computer systems. For this application, this outweighs the drawbacks: no object-oriented features, possible memory leaks using the primitive memory allocation system, and the danger of name-space collisions. Another consideration was the availability of many useful software library packages and utilities in the C language or targeted towards C, which could be useful. For example, the bison utility from the GNU project and the JPEG library from the Independent JPEG Group.

As part of the reimplementation, the routines that create the graphical objects were clearly separated from the routines that output the objects in a given format. This separation should facilitate future implementation of novel output formats. Also, it makes it easier to ensure that the appearance of the graphical objects is as similar as possible in the different output formats.

Undocumented features

There may be some undocumented commands in the current version of MolScript. These are highly experimental in nature. They may change drastically, or be removed, in the next version. It is a waste of time to try to figure out what they are and what they do.

Acknowledgements

Ethan Merritt, Mats Dahlberg and Jonas Uppenberg have contributed to the bug fixes and modifications in v2.0.1, v2.0.2 and v2.1.

For input and support during the work on version 2.0 of MolScript, I wish to thank Anders Lillienau, Mats Dahlberg, Björn O. Nilsson, Johan Kördel, Michael Åsman, Mats Kihlén, Henrik Fridén, Derek Ogg, Jonas Uppenberg and others at the Structural Chemistry department and other departments of Discovery Research at Pharmacia & Upjohn Inc., Sweden. Robert Esnouf provided input for the separate residue colour feature through his BobScript, a modified version of MolScript.

For input and support during the work on versions 1.0 to 1.4 of MolScript, I thank T. Alwyn Jones, Mats Kihlén, Ylva Lindqvist, Erling Wikman, Hans Eklund, Carl-Ivar Brändén and others at the Department of Molecular Biology, BMC, Uppsala University. This work was supported in part by Nordisk Industrifond and The Swedish Natural Science Research Council (NFR).

Eric Fauman, Michael Sutcliffe, Paul McLaughlin, Leo Caves and Arne Elofsson have contributed bug discoveries, fixes and other suggestions for versions 1.1 - 1.4 of the program.

Ethan A. Merritt wrote the interface to Raster3D in version 1.4 of MolScript.

Top page