|README.probe||Home Page: Richardsons' Laboratory||J. Michael Word - 7/2001|
The program Probe generates "contact dots" at points on the van der Waals surface of atoms which are in close proximity to other atoms ; reading atomic coordinates in protein databank (PDB) format files and writing color-coded dot lists (spikes where atoms clash) for inclusion in a kinemage.
Directly based on the "sp" program by Zalis and Richardson (following the work of Connolly), the approach is to place a small probe (typically of radius 0.25 Å) at points along the van der Waals surface of a selected set of atoms and determine if this probe also contacts atoms within a second "target" set. A flexible method for selecting the source and target atoms is available along with command line flags for altering the probe radius and dot density. Although Probe can generate "surface dots" were there are no nearby atoms, its primary use is to analyze atomic packing. For packing analysis and structure validation, Probe can generate contact surfaces within a set of atoms ("SELF dots").
For meaningful use of Probe in the study of molecular structures, coordinates for all hydrogen atoms must be included in the model. Modeling with "implicit hydrogens" is inadequate since the vast majority of steric interactions which constrain conformational choices take place among hydrogens. A program called Reduce, also available from the Richardson lab, uses simple geometric considerations to add hydrogens to a PDB file and optimize their orientations .
Probe has many options which modify the way output is formatted. Instead of kinemage format, it can write graphical information in O or XtalView format. It can calculate a table of dot information with contact score values and percent dot coverage. Finally, it can produce a detailed "unformatted" description of each dot, including source and target atom names, distances, atom types, and partial scores. Because Probe is very flexible, it is helpful to develop a working knowledge of options and especially selection criteria.
||Buttons are no longer generated for each element type by default. To generate these buttons use the
||Selections can now refer to negative residue numbers. (Sometimes you need to include extra parentheses or a space to prevent the selection from being treated as a command line option.)|
Probe was designed for UNIX and the commands described below follow the UNIX conventions. For a brief description of Probe features, run Probe without any options. The command "probe -h" will give a more complete description of program options.
In its most basic form, the syntax is:
which will generate SELF dots for all atoms in the input file except alternate (e.g., B or C) conformations and append them to the end of the kinemage file. Note the ">>" redirection symbol which stands for APPEND; in the normal case, Prekin would be used to make a kinemage of the molecular structure and Probe would be used to append the dot information.
A more extensive set of command line options is available in the format
(the parts in square brackets may be optional; by default the results go to standard output).
There are four modes set by command line flags
probe -SELF "pattern1" inputfiles >> kinfile
|### Intersect 1->1|
probe -BOTH "pattern1" "pattern2" inputfiles >> kinfile
|### Intersect 1->2 and 2->1|
probe -ONCE "pattern1" "pattern2" inputfiles >> kinfile
|### Intersect 1->2|
probe -OUT "pattern1" inputfiles >> kinfile
|### External surface|
(How the selected atoms interact is listed above as a comment after the hash mark.)
By default, HET groups and waters are included in the dot calculations but *NOT* mainchain to mainchain interactions. These settings may be changed with the
The use of patterns to specify the interaction being examined is illustrated with the following examples:
calculates self packing in all atoms from the file 1filH.pdb with a temperature factor less than 40 and an alternate conformation code of blank or "A" (-self is the default and the pattern is in quotes because it contains a space). This is a useful pattern for validating a structure because it ignores atoms which may have poorly determined coordinates. In other situations, the pattern could be replaced with "all" to select all the atoms.
To identify the interface between chain E and chain I in the file enzH.pdb
To create a table of contact statistics use -count
The example also shows the use of a single '>' mark; the UNIX signal to overwrite (!) rather than append to the output file.
Even more dot information for each dot can be tabulated with -unformated
You can create surface dots
These dots are equivalent to the non-reentrant part of a Connolly surface. When using surface dots, it is sometimes useful to expand or contract the probe radius using the
Finally, here is a sequence of Prekin and Probe commands which can create a kinemage where each category of contact is broken down separately. The patterns used give some sense of the level of control Probe permits.
Syntax: probe input.pdb >> out.kin or: probe [flags] "src pattern" ["target pattern"] pdbfiles... >> out.kin Flags: -SElf self intersection: src > src (default) -Both intersect both ways: src <=> targ -ONce single intersection: src > targ -OUt external van der Waals surface of src (solvent contact surface) -AUTObondrot filename read and process an autobondrot file shortcuts: -SCAN0 same as: -3 -mc -self "alta blt40 ogt33" -SCAN1 same as: -3 -once "sc alta blt40 ogt33" "alta blt40 ogt65,(not water ogt33)" -SCSurface same as: -drop -rad1.4 -out "not water" -EXPOsed same as: -drop -rad1.4 -out (note: user supplies pattern) -ASurface same as: -drop -rad0.0 -add1.4 -out "not water" -ACCESS same as: -drop -rad0.0 -add1.4 -out (note: user supplies pattern) -DUMPAtominfo count the atoms in the selection: src (note that BOTH and ONCE require two patterns while OUT, SELF and DUMPATOMINFO require just one pattern) -Implicit implicit hydrogens -Explicit explicit hydrogens (default) -DEnsity# set dot density (default 16 dots/sq A) -Radius#.# set probe radius (default 0.25 A) -ADDvdw#.# offset added to Van der Waals radii (default 0.0) -SCALEvdw#.# scale factor for Van der Waals radii (default 1.0) -COSCale#.# scale C=O carbon Van der Waals radii (default 0.94) -SPike draw spike instead of dots (default) -SPike#.# set spike scale (default=0.5) -NOSpike draw only dots -HBRegular#.# max overlap for regular Hbonds(default=0.6) -HBCharged#.# max overlap for charged Hbonds(default=0.8) -Keep keep nonselected atoms (default) -DRop drop nonselected atoms -LIMit limit bump dots to max dist when kissing (default) -NOLIMit do not limit bump dots -LENs add lens keyword to kin file -NOLENs do not add lens keyword to kin file (default) -MC include mainchain->mainchain interactions -HETs include dots to non-water HET groups (default) -NOHETs exclude dots to non-water HET groups -WATers include dots to water (default) -NOWATers exclude dots to water -WAT2wat show dots between waters -DUMPH2O include water H? vectorlist in output -4H extend bond chain dot removal to 4 for H (default) -3 limit bond chain dot removal to 3 -2 limit bond chain dot removal to 2 -1 limit bond chain dot removal to 1 -IGNORE "pattern" explicit drop: ignore atoms selected by pattern -CHO#.# scale factor for CH..O Hbond score (default=0.5) -PolarH use short radii of polar hydrogens (default) -NOPolarH do not shorten radii of polar hydrogens -NOFACEhbond do not identify HBonds to aromatic faces -Name "name" specify the group name (default "dots") -Countdots produce a count of dots-not a dotlist -Unformated output raw dot info -OFORMAT output dot info formatted for display in O -XVFORMAT output dot info formatted for display in XtalView -GAPcolor color dots by gap amount (default) -ATOMcolor color dots by atom type -BASEcolor color dots by nucleic acid base type -COLORBase color dots by gap and nucleic acid base type -OUTCOLor "name" specify the point color for -OUT (default "gray") -GAPWeight# set weight for scoring gaps (default 0.25) -BUMPWeight# set relative scale for scoring bumps (default 10.0) -HBWeight# set relative scale for scoring Hbonds (default 4.0) -DIVLow#.# Division for Bump categories (default -0.4) -DIVHigh#.# Division for Contact categories (default 0.25) -KINemage add @kinemage 1 statement to top of .kin format output -NOGroup do not generate @group statement in .kin format output -ELEMent add master buttons for different elements in kin output -NOHBOUT do not output contacts for HBonds -NOCLASHOUT do not output contacts for clashes -NOVDWOUT do not output contacts for van der Waals interactions -NOTICKs do not display the residue name ticker during processing -STDBONDs assume only standard bonding patterns in standard residues -NOPARENT do not bond hydrogens based on table of parent heavy atoms -SEGID use the PDB SegID field to descriminate between residues -OLDU generate old style -u output: kissEdge2BullsEye, etc -VErbose verbose mode (default) -REFerence display reference string -Quiet quiet mode -Help show expanded help notice (includes other flags) Pattern elements: (should be put in quotes on the command line) FILE# within file # MODEL# within model # CHAINa within chain a SEGaaaa segment identifier aaaa (where _ represents blank) ALTa alternate conformation a ATOMaaaa atom name aaaa (where _ represents blank) (all 4 characters are used so H would be ATOM_H__) RESaaa residue aaa # residue # #-# residue range # res residue type by one or three letter codes ALL,PROTEIN,MC,SC,BASE,ALPHA,BETA,NITROGEN,CARBON,OXYGEN, SULFUR,PHOSPHORUS,HYDROGEN,METAL,POLAR,NONPOLAR,CHARGED, DONOR,ACCEPTOR,AROMATIC,METHYL,HET,WATER,DNA,RNA all or a subset of the atoms OLT# Occupancy less than # (integer percent) OGT# Occupancy greater than # (integer percent) BLT# B-value less than # (integer) BGT# B-value greater than # (integer) WITHIN #.# OF #.#, #.#, #.# atoms within distance from point Patterns can be combined into comma separated lists such as "trp,phe,tyr" meaning TRP or PHE or TYR. Patterns that are sepatated by blanks must all be true such as "chainb 1-5" meaning residues 1 to 5 in chain B. You can also group patterns with parenthesis, separate multiple patterns with | meaning 'or' and choose the complement with NOT as in "not file1" meaning not in file 1. An autobondrot file is similar to other PDB input files but it includes information identifying atoms subject to rotations and other transformations. Example autobondrot file fragment showing Calpha-Cbeta bond rotation and a periodic torsion penalty function for this rotation ATOM 1 CB TYR 61 34.219 17.937 4.659 1.00 0.00 bondrot:chi1:78.7: 0:359:5:33.138:18.517: 5.531:34.219:17.937: 4.659 cos:-3:60:3: ATOM 1 1HB TYR 61 34.766 18.777 4.206 1.00 0.00 ATOM 1 2HB TYR 61 34.927 17.409 5.315 1.00 0.00 ATOM 1 CG TYR 61 33.836 16.989 3.546 1.00 0.00 ... Autobondrot commands use colons to separate values Transformations: BONDROT:id:currAng:start:end:stepSz:x1:y1:z1:x2:y2:z2 TRANS: id:currpos:start:end:stepSz:x1:y1:z1:x2:y2:z2 NULL # dummy Bias functions: COS:scale:phaseOffset:frequency POLY:scale:offset:polynomialDegree CONST:value Branching: SAVE and RESTORE or "(" and ")" (e.g. to rotate each Chi and the methyls for isoleucine the sequence is: rotChi1/SAVE/rotChi2/rotCD1/RESTORE/rotCG2) Set orientation: GO:angle1:angle2:... Include files: @filename Comments: # comment text Probe: version 2.6 10/01/2001, Copyright 1996-2001, J. Michael Word
Starting with version 2.0, Probe has been extended to read specially marked up fragments of a PDB file which describe dihedral rotations as well as other transformations. The command line flag
If you don't have Probe or if you have an old copy, you can get the latest release from
The source code should compile easily on almost all UNIX like systems. Copy the "Makefile" for your system
The most common problem using Probe is specifying selection patterns. Remember self dots
Output from Probe is generally designed to be appended to the end of a kinemage file. If you just want to see the dots without creating a model first, add @kinemage 1 as the first line to the dotfile (either by hand or using Probe flag: -kin) and Mage can now display it.
We hope this helps you get started looking at molecular contact surfaces. To find the latest version of Probe, see http://kinemage.biochem.duke.edu/. A comprehensive description of the small-probe method is found in  and .
1) Word, et. al. (1999) Visualizing and Quantifying Molecular Goodness-of-Fit: Small-probe Contact Dots with Explicit Hydrogens, J. Mol. Biol. 285, 1711-1733.
2) Word, et. al. (1999) Asparagine and Glutamine: Using Hydrogen Atom Contacts in the Choice of Side-chain Amide Orientation, J. Mol. Biol. 285, 1735-1747.