Difference between revisions of "DAS/1"

Revision as of 11:43, 19 March 2010

About the DAS Protocol

The current DAS specification, DAS 1.53, was written by Lincoln Stein, Sean Eddy, and Robin Dowell and is the basis for a large number of software implementations. Around 650 public DAS data sources are currently running worldwide, with many more private data sources known to be in use. A number of websites and software applications function as DAS clients.

The specification is being actively supported, and continues to be extended in order to cater for the needs of its existing users and expand its applicability to additional arenas. For example, though originally focussed on genomic annotation, extensions have enabled DAS to be used to distribute alignment, structural and molecular interaction data. These extensions are listed in the unofficial 1.53E (extended) specification.

Note: This protocol should not be confused with that of DAS/2, an entirely separate project.

The DAS Overview provides a glossary and list of concepts. The DAS Plans page describes some details of possible future improvements to DAS.

Architecture

The Distributed Annotation System comprises three types of component:

Data source

A data source provides programmatic access to data over a network, typically the internet. Each data source contains a set of data from one provider, for example Pfam domains.

Registry

The DAS Registry lists and describes public data sources and the types of data that may be communicated using DAS. It is accessible programatically.

Server

The term DAS server is often used interchangeably with DAS source. However a server is technically a piece of software used to host one or more data sources, and like the registry can provide details of these data sources.

Client

A client consumes and integrates the data contained within one or more data sources. It may also communicate with a server or registry to obtain information about available data sources.

How to set up a DAS server

Implementation

Theoretically it is quite easy to implement a DAS server (once you know how). However, there are also some well established multi-purpose server implementations designed to be as flexible and easy to set up as possible. Many distributions contain ready made data-adaptors (e.g. for GFF files). For custom data, simple plugins can be written to quickly provide your data via DAS:

• Proserver is a well established Perl server, supports all DAS extensions and is the most heavily used.
• LDAS is an older Perl server which is easy to set up but lacks support for the latest DAS features.
• Dazzle is a well established Java server tied to the BioJava framework and support most extensions.
• MyDas is a newer, more streamlined Java server but does not yet support all extensions.

Validation

Servers should endeavour to conform strictly to the DAS data formats as this maximises compatibility across the network and minimises the maintenance required as a result of evolving client software. RelaxNG documents are now available from the DAS Registry to help validate the XML produced by your servers. These are used by the registry to help servers conform to the DAS specification before being registered.

Publishing and Discovery of DAS sources

Publishing your source in the DAS Registry allows you to advertise its availability, and is described in the DAS Registry documentation. Registered sources are also easier for users to visualise in clients that are capable of interacting with the DAS Registry, such as Ensembl, Dasty and SPICE.

Training in DAS

DAS Tutorials, talks and hackathons take place once a year at the Genome Campus UK almost every year in the DAS Workshop. The next workshop is likely to be in March 2010. Events, courses and workshops involving DAS are often listed on the Current events page.