It is important that the molecule you wish to examine has been carefully prepared. Conformation hunting is time consuming and space-eating. A few points to guide you-
a) Is the molecule structurally correct? Use 'Draw' to correct and check.
b) Are the atom types correctly allocated (Even if you have minimised the structure, they may not be correct- (See the definitions of atom types in Appendix 2). Use 'Draw' to correct and check.
c) Are the Atom Centred Charges (ACCs) Allocated? Use the 'Minimise/Charge' button for this.
d) If XEDs are present (Book 5 page 1), they must be removed and the molecule recharged with ACCs.
Activating the Conformation Hunting button brings up a new window. The main screen window remains underneath and is active. The questions in the new window guide you through the setup:
The screen molecule will be the input structure. Enter the OUTPUT file name. Several files are produced by this procedure having the same filename but different extensions (i.e. jim.xxx, where xxx is .ast, .bst etc. You might enter 'jim' here, having already written the screen molecule as 'jim.dat').
The hunter may take its starting conformers directly from the input structure, or subject the input structure to a preliminary dynamics run depending on how you answer later. Regardless of the origin of the starting conformation, it will be subjected to an exhaustive bond-twisting minimisation routine. Enter the bond-twist increment (in degrees). The default is 10 degrees - this is usually adequate but can be set bigger for quick results.
The program keeps all conformations within an energy limit above the current 'global' minimum. (The default is 20 kcals/mole. All conformers within 20 kcals/mole of the global minimum will be kept up to a maximum (2000 at present)). Enter your choice of upper limit or default by hitting 'enter'.
The bond twisting can be done an infinite number of times. The larger the number you enter here, the more conformers will be found close to the global minimum (GM). For a serious investigation, 200 twist randomisations are recommended. The default is set lower because we usually want a representative set of conformations rather than many similar structures close to the GM.
The system examines your molecule for multiple bonds and asks you if you would like to hold any (e.g. a peptide bond kept trans). Reduce the window so that you can manipulate the main screen. Add labels (Book 1) and check which bonds are being referred to. The 'holds' are not guaranteed if the dynamics is requested to generate starters for rings (see next), but the system will try.
The system will detect any flexible rings in the molecule . Under these circumstances, the hunter can initially put the molecule through very high temperature, torsional dynamics (energy placed in torsional twisting rather than atomic velocities) as mentioned above. Each iteration produced a number of broadly-based starting conformations which are likely to include all ring conformers. Enter the number of dynamics runs you think necessary to find all (or at least the important) conformation starting points. The more flexible the molecule, the more you will need. 1 or 2 are usually enough. Each iteration can produce up to five starters The default (just hit 'enter') is none.
Our own experience tells us to keep the dielectric at unity. This gives hydrogen bonds more of a chance to influence conformational preference. However, extended conformers will usually not make up the lowest energy set. If you suspect a higher dielectric would be appropriate (experimental measurements at the water/peptide interface revealed a dielectric of 8 {Science 1992}), increase the dielectric. A value beyond 10 will have little additional effect.
The questions that follow, are relevant to the advanced features of XED. They create the files which are used for Field Analysis (Section 2) and specialised Drug Design. You may want to skip them in the first instance by entering 'N'o at this point. (Fields will be discussed later).
The main objective of the XEDs is to create good quality fields around a molecule which can be used to compare molecular similarity. These fields are very poorly constructed via atom centred charges (ACCs). COSMIC has devised a means of defining electron density based on orbital distribution. Such eXtended Electron Distributions (or XEDs) allow much more accurate electrostatic fields to be created. Because the addition of XEDs markedly increases the complexity of the conformational profile of a molecule, the early stages of conformational hunting and single minimisation do not employ XEDs (do not add them before conformer hunting). The next stage of the conformation hunter adds XEDs automatically to the ACC-minimised set (.bst) and reminimised into a .cst file. Finally the program completes the exercise by adding field points into a .fst file if requested.
To add fields to all conformers, default.
The comparison procedure (XEDQMFCOM), which will be described in Section 2, is very computer intensive. The next question on screen, allows you to limit the number of conformations onto which will be put fields. The defaults are relevant for the present 4000 series of SGIs.
In the comparison stage, we are usually looking at a bunch of drugs or hormones which are active at a given enzyme or receptor. If you believe that no conformations above a certain limit (say 3kcals/mole) will ever be accommodated by your chosen receptor or enzyme, then now is the chance to test you theory and perhaps to reduce the workload by reducing the number of conformers you need. At this time, it is anyone's guess how far a receptor will go in accommodating a high energy conformer.
You can probe the surface with all sorts of neutral groups. For drug design, the default for a peptide/protein environment seems sensible.
A summary of your input will appear on screen.
Send the job to BATCH by defaulting and keep an eye on the CPU ('top' or 'osview') and your mail. A log of the process will appear in the mail when all is finished.
Note: The input to this procedure is a single molecule. It produced a number of multi-files as follows. These can be viewed in the main screen (see Book 7 - ASTRAL):
.ast contains all conformers (up to 2000) from the primary bond twisting. They are not minimised.
.bst contains the minimised conformers from the .ast file, filtered to remove duplicates to RMS 0.5.
.cst contains the .bst conformers with XEDs added, re-minimised and filtered.
.fst contains the .cst files with added fields and filtered according to the input instructions for field addition.
