Wintex - a Tcl/Tk interface to Textal

Overview

Wintex is GUI front-end to Textal. It allows you to run commands by filling in fields and starting tasks. It also provides rudimentary job management by keeping files from each run in separate directories, which can be browsed. Wintex is written in Tcl/Tk, a common scripting language available on most platforms.

Here is a screen-shot example of what the interface looks like...

Requirements

Extended Tcl/Tk needs to be installed (i.e. tclx, wishx) (tclx.sourceforget.net). This is like regular Tcl/Tk, but adds some features which Wintex uses/needs.

Prior to running Wintex, you should source txenv.csh as you normally would for running Textal (see Installation notes?), to set up some environment variables. Among other things, this also sets a critical environment variable called $WINTEX, which identifies the location of the Wintex Tcl scripts. It also sets up an alias for 'wintex', making it easy to run from the command line.

How to Use the Interface: 3 Panels on the Main window

The basic mode of interaction is that you will start a job by selecting what you want to run from a pull-down menu and filling in fields of filenames and input parameters in a popup dialog box. Then you submit the job, and it starts running. It will create a sub-directory called texjobN, where N will start at 1 and increase incrementally. All the files from the current run will go in that directory, including temp files, log files, output files, and even the parameters you input and the script that got generated and executed. You will be able to track the progress of your script in a popup window which echos the output dynamically.

Back in the main window, hitting 'Refresh' will show a new line in Jobs Panel that has been created for your job, with status 'running'. Eventually, it will change to 'complete' or 'failed' (you have to keep hitting Refresh to get updates). When you click on the line for a given job, all the files are shown in the Files Panel. When you click on these, it will bring up the contents of the file in a window for browsing. The files are by time, with the most recently generated/modified files at the top.

Some important files are:

texjob.settings - contains parameters originally filled in
texjob.dat - contains title, process id, etc.
texjob.notes - contains content of annotations made by user
texjob.script - contains the actual script that was built and executed
texjob.out - contains the text output from running the script

The Notes Panel (in the middle of the main window) is an editable window for annotating jobs. For example, you might want to type in some comments about why you ran that job, how it differed from another run, and whether or not it worked. You do not have to explictly save these; you can edit them any time, and they will be saved automatically.

Back in the Jobs Panel, if you right-click on a texjob line, you can edit its title, kill the job, or delete the job. Multiple jobs can be run at the same time.

Menu Selections

Main

Set Global Defaults - see section below
Documentation - launches help pages in a Netscape browser
Quit

Model-Building - pre-compiled processes of multiple steps; just input your map and go 1) Build complete model (TEXTAL) 2) Build backbone C-alpha chains (CAPRA) 3) Build side-chains given C-alpha chains Build side-chains is an option to start the modeling process by inputting an existing C-alpha backbone. The sequence identities can be optionally specified, to force TEXTAL to use a user defined choice. If the residues are left as UNK, then TEXTAL by default will choose the identities. Build side-chains given a C-alpha chain, will automatically run lookup, followed by real-space refinement (sequence alignment is NOT used in this mode!)

Capra - input a map, get out chains of C-alpha's
Textal - input a map, get out complete models with side-chains

Individual Steps - for those who want to run specific steps on specific inputs, or tweak some parameters

Scale
Trace
Calculate features
build C-alpha chains
Stitch
refine C-alphas
Lookup
Sequence Alignment
Real-Space Refinement

Utilities

Trim chains - if you have a reference model/true structure, you can throw out all the extra chains built in a map except those superimposed on a select monomer, which helps in visual inspection of Textal models
Evaluate model - if you have a reference model/true structure, you can compare a Textal model against it to get a sense of accuracy (e.g. RMSD)
Map info - prints out some statistics and histogram of density values
Linearize trace - Outputs a "backbone-like" trace of the electron density

Global Defaults

Setting global defaults is useful but not essential. You can run every other task with setting anything here, but many of the tasks share common inputs. For example, a number of tasks require a scaled map as input. Typically, a user would have to 'Browse' the directory of a previous texjob where capra-scale was run, and select the filename of the scaled map. More easily, you can enter this particular file as a global default. That way, it will automatically be presented as the 'global' option each time a scaled map is needed in the future.

Base Name

The symbolic base name is a special global default. Many of the Textal programs create temporary files. For convenience, these may be prefixed with a common base name, perhaps reflecting the protein being worked on, or a specific version of the dataset. For example, if the base name were 'myprotein-2.8A', Textal will create files with names like 'myprotein-2.8A-scaled.xplor' (scaled map), myprotein-2.8A-capra.pdb' (C-alpha chains) and 'myprotein-textal.pdb' (final model). These are created by default, although different names can be selected by the user if desired, in dialog boxes for starting up jobs. For obscure technical reasons, the base name can only be changed through the 'Set Global Defaults' menu selection, not during job creation. If no base name is given, 'temp' will be used by default.

Filename Selection Drop-down boxes

On most job-creation dialog boxes, there are some fields that have a drop down box next to them with three options: "Locally generated file", "Global default filename", and "Browse". Obviously, "Browse" will bring up a traditional file brower dialog box, which you may use to surf to other directories and specify the file you want. "Global default filename" will fill in whatever is given as the default file to use for that parameter type, if previously specified by the user. "Locally generated file" causes TEXTAL^TM use a default filename for that parameter, assumed to be generated by an earlier step and available in the same texjob directory. Take Scale and Trace as an example. These are really two steps. The first runs capra-scale on a map input by the user (perhaps specified via Browse). If the user chooses, Textal will create an output file called -scaled.xplor, where is the base name selected by the user. This will and stored in the local texjob directory created for this job. Then capra-trace will run. It takes a scaled map as input. If the user choose, by default, capra-scale will assume the input map name is -scaled.xplor. This system of filename defaults makes it easy to script together a complex sequence of steps without have to make up and coordinate unique filenames for passing data between each step.

Texjob Directories and Accessing Output Files

As mentioned above, Wintex creates a separate sub-directory for each job that is run. After a job is complete, you may kill the GUI and retrieve files from those directories. For example, if Textal has created a model for you called 'myprotein-textal.pdb' in texjob12, then you can cd into texjob12/ and view 'myprotein-textal.pdb' in a molecular modeling program of your choice, or copy it out of there. The script file, temp files, log files, and everything else are in there too.