INTERCHEM is a general purpose molecular modelling program written by two chemists, Drs. Robin Breckenridge and Peter Bladon, at the University of Strathclyde. The original version which was mounted on Digital Equipment VAX processors driving graphics screens on a range of terminals, has now been re- written for use on Silicon Graphics workstations, where the superior graphics and computational performance have been exploited.

Throughout the design of the program, attention has been devoted to making it easy to use by relatively inexperienced people. In the SGI version of the program, only minimal knowledge of UNIX is required. It is thus possible for beginners to obtain useful results without extensive training.

The aim has been to provide a program which will be of use to chemists of all sorts; and to facilitate access to many of the well known packages (e.g. MOPAC) which are available for computational chemistry.

The popularity and usefulness of the original version of the program at Strathclyde University can be gauged by the over 10,000 instances of its use over the past 6 years.

The Silicon Graphics version of the program consists of 54,000 lines of FORTRAN code. The auxiliary program PROTEINS used to access the Brookhaven data bank consists of a further 6500 lines. In addition there are 1.2 Mbytes of data files.

A listing of the main features of INTERCHEM is as follows.

DISPLAY MODES FOR STRUCTURES (simple modes)

• Wire frame display.
• Atom-numbered wire frame display.
• Wire frame display with atom-coded half bonds.
• Wire frame display with coloured chains.
• Red-green stereo wire frame display.

DISPLAY MODES FOR STRUCTURES (with lighting model)

• CPK mode display.
• Atom-coded ball and wire frame display.
• Cylindrical bonds colour coded for bond-order.
• Cylindrical bonds with atom-coded half-bonds.

SINGLE, DUAL, QUADRUPLE DISPLAY MODES

• Display of a single structure, or of two or four structures simultaneously. (Structures labelled A, B, C, and D)

MANIPULATION OF STRUCTURES

• Rotation on X, Y, Z, screen axes
• Increase/decrease of size of structure
• Display/hide hydrogen atoms

For display modes not involving the lighting effects, manipulations occur in real time, even for structures having large numbers of bonds and atoms. With structures involving lighting models, rotations and scalings occur at acceptable rates, but the rate is dependent on the number of atoms.

OPERATIONS ON STRUCTURES

• Centre display on specified atom
• Centre display on centre of molecule
• Move structure on x, y, or z axes of screen
• Calculate bond-lengths
• Calculate non-bonded distances
• Calculate bond angles
• Calculate torsion angles
• Write file of bond lengths, bond angles, and torsion angles
• Write listing file of co-ordinates and connectivities
• Identify chiral centres
• Find molecular formula and molecular mass
• Invert structure about centroid
• Reflect structure in XY, XZ, or YZ plane
• Rotation of chain segment
• Loci of atoms on rotation of chain segment
• View down specified bond
• Remove hydrogens from structure
• Manually renumber the structure
• Re-number structure using Morgan's algorithm
• Find molecular formula, molecular weight, and masses of molecular ions in mass spectrum.

INVESTIGATION OF ENERGETICS OF BOND ROTATION

• Energy profile for rotation of a fragment about one bond
• Energy profile for rotation of a fragment about two bonds
• Energy map for rotation of a fragment about two bonds

STRUCTURE FITTING AND COMPARISON
Fitting of any of the structures, A, B, C, D on one of the others by three methods:-

• Three point fit
• Least-squares fit
• Least-squares fit with fragment rotation

STRUCTURE EXCHANGE AND COPYING
Any one of the structures A, B, C, D may be copied from any of the others.
Any pair of structures chosen from A, B, C, and D may be exchanged

STRUCTURE MERGING

The pairs of structures A and B, or C and D, may be merged to give a composite structure.

Full control of the process is achieved interactively on the screen, allowing the second (mobile) structure to be positioned and oriented relative to the first structure.

Indication is given to the operator of unacceptable non- bonded interactions.

Merging can be controlled so that a specified distance between atoms in the two structures may be met.

STRUCTURE BUILDING
New structures can be modelled interactively starting from a comprehensive library of fragments.
A previously made structure can be recalled and further modified.
The current structure can be stored away at any time.

Building takes place by adding to the current (base) structure a chosen fragment. Two basic operations are provided:
Form one bond between base structure and fragment
Form two bonds between base structure and fragment (i.e. make a ring)

Other facilities provided include:-

• Invert base or fragment structure
• Reflect base or fragment structure in XY, XZ, or YZ planes
• Alter atom(s) in base structure
• Delete atom(s) in base or fragment structure
• Alter bond(s) in base structure
• Delete bond(s) in base structure
• Form bond(s) in base structure
• Remove all hydrogens from base structure
• Add hydrogens to base structure
• Create dummy atoms in base structure
• Copy base structure to fragment area
• Exchange base structure and fragment
• Renumber base structure
• Rotate a mobile side chain in either base structure or fragment

The option of undoing the last operation is provided, so allowing mistakes to be rectified.

At any stage in the building process, the current base structure can be optimised using the molecular mechanics program PiFF. This process usually takes only a few seconds. The optimised structure is then displayed, and further extension can be carried out.

STORAGE OF STRUCTURES

INTERCHEM has a defined format for storage of structures, which is used to store both small and large molecules, including the fragments in the fragment catalogue. This format is the same as in the original VAX version of the program, meaning that data files may be transferred between the two systems.

INTERFACES TO OTHER PROGRAMS
Structures produced by INTERCHEM can be stored in files with formats suitable for submission to other programs:-

• The molecular orbital programs MOPAC, AMPAC
• The molecular mechanics program PiFF
• The distance geometry program DGEOM
• The structural data output from these programs may be read back into INTERCHEM
• Solvent-accessible surface displays using Connelly's program may be produced from INTERCHEM structures, using a version of the program bound into INTERCHEM.

SHAPES AND POTENTIAL OPERATIONS

A facility is provided to display the accessible surfaces of molecules coded to show the potentials at the surface due to charges on the atoms.

The 3-dimensional isopotential surfaces due to the atomic charges in molecules may be drawn, and potential contour maps may also be produced.

It is possible to align two molecules, without reference to their molecular structures, using as matching parameters, either the overall shape of the molecules, or the potentials on surfaces surrounding the molecules. (Method of Icosahedral Matching).

The potentials surrounding molecules due to the charges on the atoms may be displayed projected onto transparent encompassing spheres or ellipsoids.

SPECIAL FACILITIES FOR EXAMINING PROTEIN AND NUCLEIC ACID STRUCTURES

Besides the normal display methods which are available for all of these large structures (insofar as they are useful), there are special functions available. These include:-

• Colour coded wire frame display to show amino-acid or nucleotide components.
• Display to highlight structural features of proteins and nucleic acids.
• Colour coded wire frame display to differentiate protein chains.
• Display of protein structures as ribbons (also in red-green stereo).

There are functions to analyse the data contained in protein structure files in special ways:-

• Torsion angles analysis.
• Ramachandran and Balasubramanian plots to show protein secondary structure.
• A method to find significant sites in a protein structure.
• Protein structures may be truncated selectively, by selecting residues, or by retaining all of the structure within a defined radius.

Methods are provided to predict protein secondary structure features starting from the amino-acid sequence; four established methods are used, and the results from all of these may be compared. The predictions may also be edited by the user, and the resultant consensus preserved in a torsion angle file.

These torsion angle files can be used to build peptide sequences into standard INTERCHEM structures automatically. There is a choice of preparing backbone only structures, structures containing all atoms except hydrogen, or full structures containing all hydrogens. Peptide structures containing hydrogens can be made in un-ionised form, in zwitterion form, or in forms expected at different pH values. Provision is made for selectively forming S-S linkages, and also for making cyclic peptides. The structures containing hydrogen may be submitted for structure optimisation using the program PiFF directly. Alternatively, structures may be submitted to the distance geometry program DGEOM via special files.

A comprehensive package for the interactive investigation of protein homology is provided. A maximum of 60 sequences containing up to 400 amino-acid residues may be compared. Pairs of sequences may be aligned using an implementation of the Needleman-Wunsch algorithm. In addition manual alignment of whole blocks of sequences is catered for. Full 'spread-sheet' type operations are also provided; sequences may be moved, copied, edited, and stored. Sequences may be randomised (to test for significance of homology), or reversed (to test for sequence palindromy). A method is provided for examining evolutionary relationships between groups of highly homologous sequences. This searches for (apparently) simultaneous changes in residues in sequences, to find regions in a protein which may be of importance to the function of the protein.

Facilities are provided to translate DNA and RNA sequences into protein amino-acid sequences. Choice of reading frame is provided, and in survey mode all three reading frames are tried, and a report of the maximal peptide length is given. The complementary nucleic acid strand may be investigated as well.

If a version of the Brookhaven Protein Data Bank is available on the machine (or via a network), then the auxiliary program PROTEINS can be used from within INTERCHEM, to translate the data into files in standard INTERCHEM format. PROTEINS may also be used from a terminal connected directly to the PERSONAL IRIS, or from terminals connected to it via a network.

SPECIAL FEATURES FOR THE EXAMINATION OF CRYSTAL STRUCTURES

Data files from the Cambridge Crystallographic Data Base may be loaded into the system. The symmetry operations of the appropriate space group may be applied to give the structure of a single unit cell, or alternatively a nest of 27 cells (3 x 3 x 3) may be formed. Provision is made for displaying the crystal axes, crystal (Miller) planes, and an arbitrary defined axis. This axis could (for example) an experimentally determined principal electric vector etc. A comprehensive editing facility is present which allows the cleavage planes of the crystal to be exposed. Using this and the merge facility (see above) layering of one material on a substrate may be modelled.

MISCELLANEOUS FACILITIES

• Access to UNIX operations
• Display Testcard for Camera setup

PROGRAM CONTROL

With the large number of features and functions provided in the program, its control might be thought as likely to present some problems. However, it has been designed to be easy to use. Control is achieved by display of very explicit pop-up menus. These guide the user to those operations which are permissible at each stage. Structure manipulation and some other features are controlled by the use of cursor picking on static menus, and in some cases the functions of these menus is duplicated by the function keys.

COMPUTER REQUIREMENTS

Hardware

The program is designed to run on a Silicon Graphics 4D/20G machine (Personal IRIS). There should be at least 8 MBytes of memory (but preferably 16MBytes), a 380 MByte disk, and a 125 MByte tape drive.

The graphics requirements are 24 bits colour planes (used as 2 x 12 bits, i.e. double buffering is used), 24 bits Z-buffering, and 8 bit planes for cursor and over/underlay. A 19 inch 1280 x 1024 resolution screen is needed. These should be regarded as the minimum configuration for the hardware, for the program to function correctly. It will run on the later and faster versions of the Personal Iris (4D/25 and 4D/35), on models with the turbo-graphics option, and on models in the Indigo and Crimson series. It is NOT suitable for the basic Indigo machine however.

Software

The current version of the program is designed to use versions 3.3.x and 4.0.x of the IRIX operating system. In order to compile the program on the host machine, the UNIX f77 (FORTRAN) and cc (C) compilers must be available.

MANUAL

INTERCHEM is fully documented in a 250 page manual. A tutorial workbook is planned for the near future.

AVAILABILITY

The INTERCHEM system comprising INTERCHEM, PROTEINS, and PIFF is available under license through:

The license fee for industrial and government users is (US) $2000
The license fee for academic users is (US) $400

A new version of the INTERCHEM molecular modelling software is now available from QCPE and direct from Interprobe Chemical Services.

For more details take a look at the website:- http://interchem.chem.strath.ac.uk/inter/interprobe.html