Documentation for the HOLE: 3.0 Control Cards

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

3.0 Control of HOLE

When HOLE is run the variables which the user can/must specify are read from the FORTRAN input stream number 5 (defaults to the keyboard). The text output is to stream 6 (normally the screen). The i/o streams should be redirected in the manner normal for the operating system used. For unix machines this is achieved by typing the command:

(nohup nice) hole < control_file.inp > log_file (&)

where hole refers to the executable, your $PATH variable should have been modified so as point to this (see setup section>). The > and < symbols reassign the standard input and output steams to the files specified. control_file.inp is a file containing input information in the format described below and log_file.out is a file which will be created containing text output describing how the run is progressing. (The nice and & put the program into the background with lowered priority). The input is keyworded by the cards described below. In general input is case-insensitive, except for filenames - whose case is not changed on reading. Comments must be preceded by a ! and are not read. Comments can be placed at the end of any line (they will be striped off before the line is interpreted). The user is told of any non-blank line read and not recognized.

3.1 CARDS which MUST be specified:

COORD filename
This specifies the name of co-ordinate file to be used. This must be in Brookhaven protein databank format or something closely approximating this. Both ATOM and HETATM records are read. Note that if water molecules or ions are present in the channel these can be ignored on read by the use of an IGNORE card.

A new feature (in release 2.1) was the option to include a wild card (*) in the filename. e.g., coord ab*.pdb will apply hole to all files in the directory whose name starts with ab and ends with .pdb. This intended to aid the analysis of multiple copies of the same molecule - produced during molecular dynamics or other method. The hole procedure will be applied to each file in turn with the same setup conditions (initial point, sampling distance etc.). Graphics files will contain a combination of the individual runs, one after another. Note that the pdb files are read independently so that they need not have an identical number of atoms or atom order etc. (though they should be sufficiently similar for a HOLE run from identical starting conditions to be useful). It may be useful to cut down the information written out to the text output by using a SHORTO card (perhaps with level 2) if multiple files are used. Note that an alternative way to load in multiple files is a direct read from a CHARMm binary dynamics coordinate file - using a CHARMD card.

RADIUS filename
This specifies the name for the file specifying van der Waals radii for each atom. A number of files with different values are supplied with HOLE, see section 5.0. Note that HOLE unlike many programs now knows what a ~ means in unix e.g., the to one can specify a radius file in user mary's directory:
radius ~mary/hole2/rad/simple.rad

3.2 CARDS which may be specified

The most useful cards:

CPOINT x y z (where x y and z are real numbers - separated by spaces)
This specifies a point which lies within the channel, for simple channels such as gramicidin results do not show great sensitivity to the exact point taken. An easy way to produce an initial point is to use molecular graphics to find two atoms which lie either side of the pore and to average their co-ordinates. Or if the channel structure contains water molecules or counter ions then take the coordinates of one of these (and use an IGNORE card to ignore in the pore radius calculation).

If this card is not specified then HOLE now (from 2.2) attempts to make a guess where the channel will be. The procedure assumes the channel is reasonably symmetric. The initial guess on cpoint will be the centroid of all alpha carbon atoms (name 'CA' in pdb file). This is then refined by a crude grid search up to 5 angstroms from the original position. This procedure works most of the time but is clearly far from infallible - results should be careful checked (with molecular graphics) if it is used.

CVECT x y z (where x y and z are real numbers - separated by spaces)
This specifies a vector which lies in the direction of the channel/pore.

If this card is not specified then HOLE now attempts to make a guess where the channel will be. The procedure assumes the channel is reasonably symmetric. The guess will be either along the X axis (1 0 0), Y axis (0 1 0) or Z axis (0 0 1). If the structure is not aligned on one of these axis the results will clearly be approximate. Once again if a guess is used then results should be carefully checked.

SPHPDB filename
This card specifies the filename for output of the sphere centre information in pdb form. I normally use the identifier .sph for such files. The co-ordinates are set to the sphere centres and the occupancies are the sphere radii. All centres are assigned the atom name QSS and residue name SPH and the residue number is set to the storage number of the centre. The file can be imported into molecular graphics programs but are likely to be bonded together in a awful manner - as the points are very close to one another. In vmd sph objects are best displayed as "Points". Displaying .sph objects rather than rendered or dot surfaces can be useful to analyze the distance of particular atoms from the sphere-centre line.

Most usefully .sph files can be used to produce molecular graphical output from a hole run. This is achieved by using the sph_process program to read the .sph file. (Previously HOLE could directly produce surface information but this is being phased out.)

ENDRAD real number
This card can be used to specify the radius above which the program regards a result as an indicating that the end of the pore has been reached. The default value is 15.0 angstroms. This may need to be increased for large channels or reduced for small.

CONN or options can be set by: CONN Rprobe or CONN Rprobe Dconn
This turns on the Connolly option. This option was introduced in release 2.2 of hole. For each plane orthogonal to the channel vector cvect the normal hole process is performed in which the radius of a single probe sphere is maximized on ther plane. The Connolly procedure then extends this to find the additional space accessible on the plane for a probe of radius Rprobe starting from the normal hole final position. The surface produced by the HOLE Connolly should be essentially the same as locally accessible part of the MSMS or the surf procedure in vmd. The advantage of HOLE is that the external parts of the surface are cut away and that a measure is made of the accessible area for each plane.

Brief description of the algorithm: The Connolly procedure in HOLE starts from the spherical probe position. The procedure then constructs a 2D grid in the plane with a spacing of Dconn. The procedure then proceeds to find points on the grid that are adajacent to accepted points that can be accomodate spheres of radius Rprobe or above. The points are then added to accepted list. When no more points are added to the grid the edge of the accepted area is then adjusted so that the radius exactly matches Rprobe.

It should be noted that the Connolly procedure takes longer than normal HOLE and it also results in many more sphere centres in the SPHPDB file. The default value for Rprobe is 1.15 Å - this is the radius necessary to accomodate a water molecule with the "simple.rad" set of van der Waals radii. The default value for Dconn = 0.7*Rprobe. The only way to get molecular graphics output from the CONN procedure is via SPHPDB output.

*** Should add a longer description of algorithm and results that can be obtained with CONN ***

IGNORE list_of_3_character_residue_types
For fairly obvious reasons, in general it is a good idea to not consider ions or water molecules within an ion channel during a HOLE run. In the bad old days the only way to this was to edit out these records from the input file (specified by a COORD card). Now this can be achieved by specifying an IGNORE card followed by a list of residue types (which are three character strings e.g., alanine = ALA). Any atom whose residue name matches part of the list will be not be read in. Note that wildcards are not allowed in the match (if this is a pressing problem tell me as this could be changed).

To ignore all waters a line like:

ignore hoh tip wat

should appear in the control input file.

Other cards:

SHORTO (optional integer)
Normally during a HOLE run information as to the progress of the run is continuously written to the output stream (which is normally reassigned to a file). It is informative to look at this information if one is doing an initial run as it can reveal such things as whether a sufficiently large number of Monte Carlo steps/step size has been specified. But if one is performing many runs (e.g., repeated runs on a molecular dynamics trajectory) this information just takes up space. By specifying a SHORTO card the amount of information written out can be reduced - the optional integer controlling how much or little is written:

SHORTO 0
Full text output (equivalent of not specifying card!)
SHORTO 1
All text output given except "run in progress" (i.e., detailed contemporary description of what HOLE is doing).
SHORTO 2
Ditto plus no graph type output - only leaving minimum radius and conductance calculations.
SHORTO 3
All text output other than input card mirroring and error messages turned off.

SAMPLE real_number
Specifies the distance between the planes used in the HOLE procedure. The default value is 0.25 Å, this should be reasonable for most purposes. However, if you wish to visualize a very tight constriction then specify a smaller value.

MCDISP real_number
Specifies the maximum step length to be taken in the Monte Carlo Simulated annealing The default value 0.1 Å. For better sampling you can set this number higher: a good starting point is 1/10th of the minimum radius of the channel.

MCKT real_number
Specifies the initial "Boltzmann factor" to be used by Monte Carlo routine. The default value 0.1 Å. The higher the value the more likely that the routine will jump from any local minimum. Use caution as very high values will produce jumps through the wall of the pore. Generally its a good idea to do a number of initial exploratory runs with high values of MCDISP and MCKT if you have a channel with a complex network of routes.

MCSTEP integer
This card specifies the number of steps taken for the Metropolis Monte Carlo simulated annealing optimization routine. The default value is 1000 steps which is normally adequate.

RASEED integer
This can be used to provide the random number generator used by the program an initial seed. Normally the seed number is provided by a routine from the time of day. HOLE will output the value of the seed it used. You can use this card if you want to repeat a run

CHARMD filename
Does multiple HOLE run on positions taken from CHARMm binary dynamics format .DCD trajectory file. N.B. file must have exactly the same number of atoms in exactly the same order as the pdb file specified by the COORD card. Note that if this option is used the pdb file is used as a template only - the coordinates are ignored. Note that structural parameters determined for each individual structure are written in a tagged format so that it is possible to extract the information from the text output file using a grep command. The reading of the file can be controlled by a CHARMS card.

CHARMS integer integer
Controls reading of the CHARMm binary dynamics format .DCD trajectory file specified by the CHARMD card. The first number is the number of positions to be skipped before analysis is to be performed. The second number is the number of positions to be skipped in between each analysis. e.g., Suppose your .DCD file had a position written every 0.5ps and you wanted to analyze the run every 1 ps but ignoring the first 25ps (as this is equilibriation). In this case use a line:
charms 49 2.

2DMAPS filename_root
This card is used to produce files which can be used to make 2d contour maps of the inside surface of a channel. For full details see section 6.5. Note that this card is incompatible with a CAPSULE card.

CAPSULE
Turns on the capsule option. This analyzes the anisotropy of the channel. This also turns on a empirically corrected prediction of the conductance of the channel in KCl: see Section 8 for details.

3.3 Cards which are deprecated

These cards are currently recognized by the hole program but are deprecated. They were mainly used to directly produce molecular graphics output from hole: this is now done by using SPHPDB output and the sph_process program. It is likely that these CARDS will be removed from the next (2.3) release of hole
MOLQPT filename
This facility is for people who do not use quanta but want to produce a composite picture of the HOLE surface with the molecule. If you specify a MOLQPT card then a quanta plot file will be written with a stick representation of the molecule written to colour number 1. This file can then be used with the file specified by a PLTOUT card to produce colour/monochrome hardcopy using the program qplot. Quanta users can produce more sophisticated versions such a file using the plot molecules option under the file menu.
PLTOUT filename
specifies the filename for binary 3D QUANTA plot file output. It is normal to give the file the extension ".qpt" for compatibility with other programs. The binary plot file will contain a graphical representation of the hole run. It can be directly displayed in conjuction with the molecule using the QUANTA program, or can be converted for use with other programs (see qpt_conv). If card is NOT present then no graphical output will be produced. The file will contain records consisting of 4 real numbers. If you are interested in other commands please consult Molecular Simulation documentation. The commands used in hole and ancillary programs are:

1.0 x y z
change to colour x.
Extension to standard format used by HOLE if y=-55 then z is an alternative colour number (in range 15 to twenty). The alternative colour number is used by qplot so that HOLE objects are output to different colours: see colours used in HOLE section.
2.0 x y z
move to x y z
3.0 x y z
draw to x y z from present point
4.0 x y z
dot at x y z
5.0 x y z
write character string of length x characters at the current point. This record will be immediated followed by a the character string (length x to be written).
Extension to standard format used by HOLE Program qplot uses the number z as a positioning indicator, if z=
1 the label is placed below the atom to the left
2 the label is centred below the atom
...
5 the label is centred on the atom
...
9 the label is placed above the atom to the right.
An easy way to remember what the integer means is to look at the keypad of any keyboard:
789
456
123
The use of this integer allows the production of pictures in which atom labels are not obscured by the molecule. The program labqpt facilitates the production of such labels.

The .qpt file produced by HOLE will contain a number of objects:

colour #
6 lines joining sphere surface to atoms causing constriction
4 The centre line of the pore (unless turned off by a CENOFF card).
7 Dots showing the surface of locus of the sphere as it squeezes through the channel (unless turned off by a DOTOFF card).

The numbers have been chosen so as to produce sensible colours when displayed in QUANTA.

Note that:

  • .qpt files are binary and files produced by one machine may not be correctly read on another (see qpt_conv - option 'A')
  • dots are not very visible in QUANTA (see qpt_conv - option 'C')
  • to avoid the final surface appearing like a bone or dumb bell dots are not draw for the first and last sphere centres determined.
  • the file can be converted for display with other programs (see qpt_conv).

DOTDEN integer
This number controls the density of dots which will be used by the program. A sphere of dots is placed on each centre determined in the Monte Carlo procedure. Only dots which do not lie within any other sphere are considered. The actual number of dots written if therefore controlled by dotden and sample. Dotden should be set to between 5 (few dots per sphere) and 35 (large number of dots per sphere). The default value is 10.

DOTOFF
This card will turn off the drawing of a dot surface representation of the hole. If not present then dot surface will be drawn.

SPIKES
The presence of this keyword will cause the program to produce spikes from the surfaces of spheres which go out until line hits an atom's surface rather than dots. This is intended to be useful in looking at channels which are not locally cylindrically symmetrical.

LINON
This will turn on the drawing of lines joining each sphere surface to the two closest atoms. This can be use in showing which atoms cause constrictions in an ion channel. In the previous version of HOLE the default option was to draw closest atom lines and a LINOFF card was specified not to do so - such a card is now ignored.

CENOFF
The presence of this flag will turn off the drawing of the locus of sphere centres as the HOLE sphere squeezes through the channel. This line can be regarded as the centre line for the channel. If a CENOFF card is not present the centre line will be output to the quanta plot file specified by the PLTOUT card.

STOP
Input of control data is stopped by reading an end of file or a "stop" card. This card is included because in the good old days of VAX/VMS this was necessary.


 <<Previous Section  Hole Doc. Index^  HOLE home  Next Section>>
Copyright 1997, 2004 by Oliver S. Smart