Documentation for the HOLE: 6.0 Visualization of HOLE results
Copyright 1997 by Oliver S. Smart

 <<Previous Section  Hole Doc. Index^  HOLE home  OSS home  comments?  Next Section>>

Navigation:

Back to contents
Back to previous section 5.0 van der Waals radii files
On to next section 7.0 Ancillary programs

6.0 Visualization of results

6.1 Plotting 2D Graphs

***to be updated***

One of the more useful ways to visualize the results of HOLE is to plot a graph (all those school teachers/university demonstrators must have some influence). Raw data which can be used for this purpose is written at the end of the text output file. This can be read into your favourite graph plotting program (I use sigmaplot). Clearly the pore radius determined should be placed on the y axis. For the x-axis you can either use the distance moved along the pore centre-line from the initial point (the first column of the data) or the x, y or z co-ordinates of the sphere centre (if the pore lies approximately in one of these directions). The former representation gives an indication of the straightness of the pore but comparison between runs is made more difficult by side to side jumps in the centre line. The latter representation, which was suggested by Mark Samson, is probably preferable as it allows easy comparison between the results of different runs and for the position of important atoms/residues to be marked on the graph.

6.2 Using molecular graphics

* Quanta
The first step is to produce a .qpt file of a HOLE run using a PLTOUT card in hole, or using the program sphqpt with a .sph file produced by a SPHPDB card in HOLE (the hole surface can be produced as a grid and colour coded using this option.

The QUANTA 3D binary file .qpt can be displayed using QUANTA (or if you have ancient E & S ps300's its predecessor HYDRA). The procedure for this is first to import the pdb file of your channel into the QUANTA (so that you can see it). Then use your mouse to select the option "Create objects" (under the heading draw). You are then confronted with a "dialogue box" asking for what type of object: select "quanta plot file" and a name for the object. The molecule and the hole through will then be displayed in conjunction.

You may find that the dots of the dot surface do not show up very well in QUANTA. If this is the case you can use program qpt_conv (see section 7.1) to convert the dots to small crosses (a new .qpt file is produced).

* Sybyl
The first step is to produce a .qpt file of a HOLE run using a PLTOUT card in hole, or using the program sphqpt with a .sph file produced by a SPHPDB card in HOLE (the hole surface can be produced as a grid and colour coded using this option.

The .qpt file must then be converted into a form suitable for use with sybyl. Use program qpt_conv (with the option "S") to produce .plt2 file. To use this file within sybyl simply display your molecule without centering on input and then at the Sybyl> prompt type:

Sybyl> take filename.plt2 
This works with sybyl6.0. At Birkbeck with sybyl6.1 the program crashes but this is a problem with the program as it occurs even with plot file produce using sybyl!

You may find that the dots of the dot surface do not show up very well in QUANTA. If this is the case you can use program qpt_conv (see section 7.1) to convert the dots to small crosses (a new .qpt file is produced).

* O
O is the standard crystallographic refinement package. It runs on Silicon Graphics, Hewlett Packard and Evans & Sutherland workstations and is available at reasonable cost from Prof. Alwyn Jones (<alwyn@xray.bmc.uu.se>), for further info see the O home page.

The first step is to produce a .qpt file of a HOLE run using a PLTOUT card in hole, or using the program sphqpt with a .sph file produced by a SPHPDB card in HOLE (the hole surface can be produced as a grid and colour coded using this option.

The .qpt file must then be converted into a form suitable for use with O. Use program qpt_conv (with the option "O") to produce a "draw_object" file which I give the extension of .obo. This is how I got to display the example molecule and the the converted output of sphqpt (run in the ~/hole2/example directory):

  O > s_a_i
 Sam> Name of input file: 1grm_single.pdb
 Sam> O associated molecule name: gram
 Sam> File type is PDB
 Sam>  Nothing marked for deletion, so no compression.
 Sam> Space for    142656 atoms
 Sam> Space for     10000 residues
 Sam> Molecule GRAM contained 34 residues and 272 atoms
 Sam> Centre of gravity updated for     1   34
  O > mol
  O >  Current molecule  has not been loaded.
 Mol> Molecule code name []: gram
  O > zone
 Mol> Zone [all molecule]: 
  O > end
  O > draw_object example_grid.obo 
 As3> O descriptor in computer file system
  O > 

You may find that the dots of the dot surface do not show up very well in QUANTA. If this is the case you can use program qpt_conv (see section 7.1) to convert the dots to small crosses (a new .qpt file is produced).

* mage
Kinemage is a protein display and viewing language developed by David Richardson. Kinemages can be viewed by the mage program which is available in versions for MS-windows, mac and various unix machines. Kinemages can be prepared from a pdb file by using the prekin program. Both programs are public domain and very easy to use. If you do not have access to a "proper" molecular graphics program or want to use HOLE on a low performance machine (i.e., PC or mac) then I recommend you use this program. A page on the PPS'96 protein structure course gives details where to get copies or try David Richardson's ftp site.

The first step is to produce a .qpt file of a HOLE run using a PLTOUT card in hole, or using the program sphqpt with a .sph file produced by a SPHPDB card in HOLE (the hole surface can be produced as a grid and colour coded using this option.

The .qpt file must then be converted into a form suitable for use with mage. Use program qpt_conv (with the option "K") to produce .kin file. This file will contain kinemage instructions for the HOLE objects alone. To view the HOLE objects in conjunction with the molecule then you should use an editor to merge this file with the output of prekin (if you are doing this under ms-windows then use write rather than notepad as the editor as the files are large). To get decent rotation rates one must reduce the number of lines to be drawn on a screen - so either turn off all HOLE objects when adjusting the view or make sure you use a low dot density (set by a DOTDEN card or in sphqpt). Depending on the machine the lines drawn can be rather thick - if this is the case try turning on the "thin line" option in the "Options menu. Note that mage can produce side by side stereo so if you can see stereo directly then put this option on or otherwise try to get a pair of mirror type stereo glasses.

If you have configured your web browser (if you have this problem look at this guide from Protein Science) to be able to view kinemages the have a look at this (30kb) example.

* InsightII
The first step is to produce a .qpt file of a HOLE run using a PLTOUT card in hole, or using the program sphqpt with a .sph file produced by a SPHPDB card in HOLE (the hole surface can be produced as a grid and colour coded using this option.

The .qpt file must then be converted into a form suitable for use with InsightII. Use program qpt_conv with the option "I". Unfortunately I no longer have access to a copy of the program so the conversion program is unaltered from release 1 of HOLE - in which there was little colour coding. Here are the notes from the previous release:

the routine writes out a Biosym Insight II special user format file (.usr). This should allow display of the dot surface and lines produced by HOLE with InsightII. There are some differences between the file formats. A single .usr file can contain only dots or lines - not both. Therefore you must run the program twice to convert a typical .qpt file produced by HOLE. Also because the .usr file format for text records essentially only allow one record per file I haven't bother to convert these. Also the control of colour is very different between the two formats. A LINE .usr file does not have any facility for changing colour in the file - all records from the .qpt file will therefore be displayed in a single colour. If you don't like this you will have to use an editor to split up the files. The DOT .usr file allows the specification of a different colour for each dot (by a number between 0 an 360). At present all dots are given the same colour which can be specified by the user. To display the dot surface (or line objects) first read in your molecule (from pdb file or similar) and then use the option "get" under the menu "user" to specify the .usr file produced by qpt_insight. The "Reference_object" option should be "on" to avoid the object being centred. From an extremely preliminary use it appears that InsightII suffers from much the same problem as QUANTA in that dots do not show up as being very bright - so it may be worth using program qpt_conv (see section 7.1) to convert the dots to small crosses (a new .qpt file is produced).

If you want to use colour coding etc. then the subroutine qptins will need updating - either does this yourself or e-mail me so that I can do it and you can test it

* rasmol *
Rasmol is a very popular molecular viewer written by Roger Sayle of Glaxo Wellcome. It is free and available in versions for pc's, mac's, unix-boxes and vms, for further information see The Rasmol Homepage. Note that to be able to use rasmol with HOLE you need to use version 2.6 (version 2.5 will not work). You can download rasmol from a number of places (see The Rasmol Homepage), for instance Edinburgh.

Unfortunately rasmol does not have any way to import graphical objects as distinct from atoms. A work around (suggested by Roger Sayle) is to treat the objects as atoms in a pdb file and use "CONECT" records to join them up. (Another annoying thing is that rasmol will rework out bonds if there is not at least one bond per atom - which means that each move draw has to be represented by 3 lines not one!) There is yet a further complication - you cannot just append the resulting "object pdb-file" to the end of the molecule's pdb file as rasmol will then work out a new bonding list based on distance (and a mess will result!). At present the only way to see the molecule and hole object is to use the MOLQPT card to produce a stick plot of the molecule and concatinate the result with the HOLE object .qpt (from HOLE or sphqpt). This can be converted into a pseudo-pdb file for rasmol. Note one loses all atom information on the molecule this way. It would be possible to be more smart about this but it may not be necessary as new versions of rasmol which can read more than one molecule in are being developed which may avoid the problem. This example (run in ~/hole2/example) shows how you can use it:

jura> cat example.qpt stick.qpt > example_and_stick.qpt
jura> qpt_conv
 *** Program qpt_conv ***
 Version H2alpha1
 (c) 1996 Oliver Smart & Birkbeck College, All rights reserved.
 Program linked at Wed Feb 26 13:33:59 GMT 1997  
 Last modified .f files: 
        8318 Feb 26 13:33 qptras.f       
        3726 Feb 24 11:45 qpt_conv.f     
       27622 Dec  5 14:15 hcapgr.f       
 
 This program converts a .qpt file (as produced by hole) 
  to something else.
 Output options
 'A' to/from ascii version of original .qpt (can then edit)
 'C' A .qpt file in which dots are replaced by 3D crosses
 'L' A .qpt file with long lines split into smaller sections
       (useful for proper depth queueing in qplot)
 'I' InsightII format
 'R' Rasmol format
 'S' Sybyl format
 'K' to David C. Richardson's kinemage format
 'O' for use with O program
 'V' to Virtual Reality Markup Language
  Enter conversion option character : r

 S/r qptras. 
   Reads in a hydra/quanta 3D binary plot file and writes
   out a pseudo-pdb file equivalent suitable for rasmol.
   
   The only way to see a hole object and the molecule which 
   produced it is to concatinate the result of a MOLQPT card
   with the object - see HOLE documentation for details.
 
 Please enter input binary hydra/quanta plot (old) filename
 defaults  ext:<.qpt> (abort by EXIT or ^D) : 
 Please enter output pseudo-pdb file (new) filename
 defaults  ext:<.pdb> (abort by EXIT or ^D) : 

STOP FORTRAN STOP qpt_conv: normal completion statement executed
jura> 
to view the file you create start rasmol (assuming you have it on your path)
jura> rasmol example_and_stick_rasmol.pdb
RasMol Molecular Renderer
Roger Sayle, August 1995
Version 2.6
[24bit version]

Number of Groups .... 1
Number of Atoms ..... 5688
Number of Bonds ..... 5688
RasMol>
If you have setup the application type correctly in netscape you can see the result by clicking this link.

* VRML - virtual reality markup language
Virtual reality markup language may prove to be very useful for molecular scientists (for details see IChem Virtual Molecular Studio and VRML in Chemistry by Horst Vollhardt). For one thing it might establish a universal way of communicating 3D objects - avoiding the need to write conversion programs (7 in HOLE so far) and to be able to do good molecular graphics on quite cheap pc's.

VRML is supported in two ways by HOLE. The first is an option in qpt_conv (V) which is similar to those described above and converts the simple 3D move/draws so that they can be used in VRML. The second is the program sph2vrml which allows the production of solid spheres instead of dot surfaces, this was developed by Xiaonan Wang, a student at Birkbeck. The program produces a pretty crude output at present - great advances are expected from more sophisticated approaches being introduce by another student Guy Coates.

(more to be added).


6.3 Colours used in HOLE

The graphical output of HOLE is in the form of a .qpt file - QUANTA binary plot format. The format of the file is described in the discussion for the
PLTOUT card. HOLE objects are divided as follows: (links are provided to sections which give information how to produce/turn off each kind of object)

Object Quanta Alternative Default Quanta Default qplot Colour # Colour # Colour colour output
Centre line 4 15 Yellow Thick yellow Normal pore surface 7 17 Bright Green Open Green Low radius surface 3 16 Red Open Red High radius surface 2 18 Light Blue Open deep blue Capsule vectors 8 19 Purple Thick Cyan Closest atom lines 10 20 Salmon Thick Grey Stick copy of molecule 5 - White White/Magenta
In general if you use a molecular graphics program other than quanta then the output of program qpt_conv should conform to these colours as much as possible.


6.4 Getting hard copy pictures

Once you have visualized the hole through your ion channel you may wish to get a hard copy representation of the results. Section *** describes program qplot which produces publication quality postscript output with options for stereo pictures and colour output.


6.5 Creating 2D Maps of the internal surface of a pore

(follow link for details).
Navigation:
Back to contents
Back to previous section 5.0 van der Waals radii files
On to next section 7.0 Ancillary programs

 <<Previous Section  Hole Doc. Index^  HOLE home  OSS home  comments?  Next Section>>