One of these omissions has just been plugged by Martin Würthner (firstname.lastname@example.org) who has ported Roger Sayle's X11 package, RasMol, to RISC OS. RasMol is a truly professional package for the representation and visualization of highly complex three dimensional molecular models. It allows the structures of molecules such as proteins and DNA, made up of thousands of individual atoms, to be viewed on screen, rotated, moved and magnified. The RasMol package is used by scientists and students to look at these structures, and also for the preparation of publication quality figures for use in lectures and reproduction in journals and text books.
The determination of molecular structures through the use of techniques such as X-ray crystallography or Nuclear Magnetic Resonance (visualization) spectroscopy, involves a lot of computer intensive data processing. Traditionally, and for obvious reasons, this intensive number crunching is carried out using mainframe computers and top end UNIX work stations. However, once the data has been computed, the results can be written out as the relative three dimensional molecular coordinates of every atom in the structure.
To allow this structural information to be exchanged, standard data formats have been adopted. One major standard in this area is the Protein Data Bank format (PDB files) as used by the Brookhaven Database of sequences, made publicly available over the Internet by the USA National Institute of Health and mirrored in the UK by the European Bioinformatics Institute (URL http://www.ebi.ac.uk/).
The Protein Data bank (PDB) is an archive of experimentally determined three dimensional structures of biological macromolecules, serving a global community of researchers, educators and students with over 4,300 structures currently available. RasMol can understand PDB files as well as other formats: 'mdl' (Molecular Design Limited's MOL format), 'alchemy' (Tripos' Alchemy), 'mol2' (Tripos' Sybyl Mol2), 'charmm' (CHARMm) and 'xyz' (MSC's XMol XYZ) produced by the different software packages. Acorn filetypes have been allocated for the PDB format, as well as some of the others, allowing files to be loaded into the application with a double-click.
RasMol (Copyright ©1992-94 by Roger Sayle) is available as a C source program designed to compile and run under X11 and optimised for compilation using the GNU C compiler, gcc. The package has also been ported to and runs on Microsoft Windows, Apple Macintosh, and VMS systems and, of course, now native under RISC OS.
The program reads in the data files and then opens up a display window in which the molecular model is shown in one of a number of different representations, such as depth-cued wireframes, 'Dreiding' sticks, spacefilling (CPK) spheres, ball and stick, solid and strand biomolecular ribbons, atom labels and dot surfaces. The model can also be coloured in various ways to highlight various chosen features of the whole molecule, or of individual atoms or residues which make up the structure.
The application opens two windows, the molecule display window and a command window. This is almost identical to the way the application appears on other platforms, so those with experience of RasMol on say UNIX or Windows, will feel immediately at home. A few commands, to enable the structure to be moved, rotated and magnified, are directly available using an attached toolbar and a combination of mouse drags (using select or adjust). Simple display features such as model representations and overall colours, are also available using the menu button and then following options offered. However, for really complex operations, it is necessary to use the commands available in the command window.
Fortunately for beginners, one of the commands available is Help, which gives access to a manual of information on the commands available, and how to use them. Using the commands, it is possible to apply different display features to selected parts of a molecule, and to label the features shown. Highly complex figures can be generated in this way, and the viewer can look at them from all angles and at many magnifications, including zooming into the centre of a molecule and looking at the atoms inside.
Clearly, this can be a laborious process to replicate, so it is possible to automate the commands by saving script files of multiple commands which will load and display the molecule in a predefined way. As the application runs on multiple platforms, it is possible to import both the molecular structures and the script files from one machine to another, and to generate similar displays.
For example, at work, I carry out complex molecular modelling on a Silicon Graphics Iris Indigo workstation, I then bring the files home and can carry on working on them using RasMol for Windows on the PC486 card on my RiscPC. Now, with this new release, I am pleased to find that I am also able to run them under native RISC OS. A nice feature of RISC OS, of course, is that it is possible to load scripts into text editors such as Edit, and to rapidly and easily carry out global changes before re-running them, all within a multitasking desktop.
I haven't carried out any detailed timings, but my impressions are that the RISC OS version (currently RasMol 2.6, RISC OS version 1.10) running on a RiscPC 600 with 2Mb VRAM in an 800×600, 32,000 colour mode feels about the same speed as RasMol 2.5 for Windows running on the PC486 SXL processor card in a similar screen mode. I have just upgraded to a PC486 DX4/100 card and this now runs RasMol faster, presumably partly as a result of the numeric coprocessor available on the new card.
However, an important feature of the RISC OS version of RasMol is that it can run on any Acorn machine with OS 3.1 or higher and with 2Mb of RAM available. I would suggest that 4Mb of RAM is more appropriate, and indeed I have obtained very good results loading a structure of 10,400 atoms into a 4Mb A5000 and also a 4Mb ARM2 A3000. This must make the package very attractive for school use where RiscPCs may not always be available.
Martin Würthner has made a very good job of porting the package across, and the colour screen output, to my mind, is better under RISC OS than under Windows. A nice feature is a dither tool available under RISC OS which, when clicked on in a 256 colour mode, generates a better dithered representation of the current view, albeit with a time penalty. The final images, once generated, can be exported in a number of formats for possible publication or incorporation in other applications. Filetypes currently supported are sprites, GIF and EPS (both bit-mapped and vector PostScript, although there are some problems at present with the latter).
To really do justice to this application, and to find out all of the many sophisticated features available, you need to get hold of a copy and to play with it. Many scientists and students will already be familiar with the package as used on other platforms, but there must also be many teachers, for example, who have never seen this application. I strongly urge them to get hold of a copy, because I know from experience that there is nothing like being able to manipulate models in three dimensions to enable a more rapid learning process of concepts in chemistry and molecular biology.
The package, as supplied on HENSA, includes some example script files, and a few simple molecules such as aspirin and atropine as well as a 'bucky ball' and a Protein DNA complex. Further structures are readily available on the Web, and by FTP, from various academic sites including EBI.