- USING GBROWSE AS A DAS SERVER & CLIENT
USING GBROWSE AS A DAS SERVER & CLIENT
The Distributed Annotation System (DAS; www.biodas.org) is a simple protocol for sharing and integrating genomic annotations. GBrowse can act both as a DAS front end for visualization and as a DAS server for sharing your annotations.
For a good introduction to the DAS features of GBrowse, see the online tutorial at http://www.gmod.org/ggb/tutorial/tutorial.html#DAS. This document is a concise reference to the DAS integration features.
You will need to install the Perl Bio::Das module before you can use any of the DAS features. This module is available from CPAN ( http://www.cpan.org ), or via the ActiveState PPM tool in the repository http://www.gmod.org/ggb/ppm .
CONFIGURING GBROWSE TO BE A DAS SERVER
At installation time, the "das" CGI script was installed in your CGI-BIN directory. This script is the core of the GBrowse DAS server. None of the GBrowse databases on your system are initially available as DAS sources until you explicitly activate them by adding the "das mapmaster" option to the configuration file. Similarly, none of the tracks in a given database are exported as DAS-available tracks until you explicitly add a "das category" option to the track configuration stanza. Together these two features allow you to select which databases will be exported as DAS-accessible databases, and which tracks within a DAS-accessible database will be visible to the outside world.
The following DAS-specific options are recognized:
- das mapmaster
This option, which should appear somewhere in the [GENERAL] section, indicates that the database should be made available as a DAS source. The value of the option corresponds to the URL of the DAS reference server for this data source, or "SELF" if this database is its own reference server. (See http://www.biodas.org/ for an explanation of what reference servers are.)
das mapmaster = SELF das mapmaster = http://www.wormbase.org/db/das/elegans
- das category
This option must appear in each of the track configuration stanzas that you wish to export as DAS-accessible data. Note that it is not sufficient to define a das category in the [TRACK DEFAULTS] section. The value of this option should be one of:
repeat transcription translation variation experimental structural miscellaneous
which correspond to "DAS annotation type categories" as explained on the DAS web site.
- das landmark
If this option appears in a track configuration stanza and is set to a true value (any non-zero value) then the DAS server will treat the features in this track as landmarks and allow remote users to search for these features by using their class and name.
- das subparts
If this option appears in a track configuration stanza and is set to a true (non-zero) value, then the DAS server will indicate that features of this type contain subparts of a genome assembly. This relationship can be used to indicate that a contig contains reads, for example. This option is only used for reference servers that serve genomic assemblies and is not well tested so should probably be avoided for now.
- das superparts
If this option appears in a track configuration stanza and is set to a true (non-zero) value, then the DAS server will indicate that features of this type are themselves superparts of a genome assembly. This relationship can be used to indicate that a read feature is contained in a contig. This option is only used for reference servers that serve genomic assemblies and is not well tested so should probably be avoided for now.
CONFIGURING GBROWSE TO BE A DAS CLIENT
GBrowse can run as a DAS client as well, either on a track-by-track basis, or by configuring GBrowse to run entirely off a DAS source. However, because of a poorly characterized interaction between the Bio::Das module and Perl 5.6, it is recommended that you use Perl 5.8.1 or higher for this. Otherwise you may experience out of memory errors.
You can display individual DAS tracks in GBrowse just by typing in the URL of a DAS server in the "Add remote annotations" section of the GBrowse main web page. The syntax of the URL is simple:
Insert all features defined in database hg16 (human genome release from July 2003) from the DAS server located at genome.cse.ucsc.edu.
Insert all features of type "genscan" from the UCSC database.
Insert all features of type "genscan" and of type "mgcFullMrna."
Insert all features of category "variation."
Insert all features of category "variation" or of type "genscan."
You may wish to use the "remote sources" option to preconfigure a popup menu of remote DAS servers that users access frequently.
Alternatively, you can configure GBrowse so that it runs entirely off a DAS server. To do this, simply set the db_adaptor and db_args options to use the Bio::Das data adaptor. For example:
[GENERAL] description = Human July 2003 Genome at UCSC db_adaptor = Bio::Das db_args = -source http://genome.cse.ucsc.edu/cgi-bin/das -dsn hg16
The db_args option expects a -source argument that points to the base of the DAS server, and a -dsn argument that points to the data source name.
You will also need to configure a track for each of the feature types served by the remote DAS server. You can get a list of such feature types simply by fetching the following form of DAS URL:
Alternatively, to make your life more convenient, GBrowse comes with a utility script called make_das_conf.pl that will create a rough-and-ready configuration file for you, given the URL and data source name of a DAS server. Instructions on using make_das_conf.pl can be found using the "perldoc" command:
% perldoc make_das_conf.pl
After generating the config file, you will probably want to customize it. Particularly in need of customization are typically the glyphs used in the track definitions, and the aggregators.
For further information, bug reports, etc, please consult the mailing lists at www.gmod.org. The main mailing list for gbrowse support is firstname.lastname@example.org.
For support with DAS, please subscribe to the biodas mailing list. Details can be found at http://www.biodas.org/
Lincoln D. Stein, email@example.com