Tutorial: An Introduction to the Basics of the DAS Protocol
By making use of DAS you can take advantage of being able to view integrated information from multiple sources, without these sources needing to be aware of each other. You can also add your own DAS data source, perhaps privately in your own institution and then view the information served from this source in the context of features from other institutions.
This tutorial will enable you to understand how DAS servers function and how you can find relevant DAS data sources.
Example DAS Reference Server – The UniProt DAS Reference Server
This section will illustrate the structure of typical DAS requests and responses, including those that you might expect to see when things go wrong - the entry point you have asked for is not available for example.
DAS information is made available from DAS servers. A DAS server is a web application that serves data in the form of a series of XML documents that can be read and processed by DAS client software. There are several different types of XML document, each having a specific function, such as:
- providing the sequence of the molecule (nucleic acid or protein);
- providing details of the features coordinated on a molecule;
- providing a summary of the features coordinated on a molecule.
If features on a specific protein are being obtained from several different locations, it is very useful to ensure that one reliable, common sequence has been obtained. To achieve this, DAS servers are separated into two kinds:
- Reference servers provide the molecule sequence.
- Annotation servers provide features upon the sequence, such as gene structures, regulatory motifs, protein domains, or related to the protein, such as journal references. Annotation servers often refer to a specific reference server as their 'map master', i.e. the server that you should expect to be able to retrieve the corresponding sequence from. This mechanism has become less important since the inception of the DAS Registry Server however.
You will now investigate an example of each kind of server.
The UniProt DAS Reference Server
The UniProt DAS reference server has the primary purpose of providing sequence information to DAS clients from the UniProt Knowledge Bank. It can be queried using UniProt accession numbers, Swiss-Prot protein Ids and also IPI (International Protein Index) accessions.
Navigate to http://www.ebi.ac.uk/das-srv/uniprot/das/dsn Here you will find a summary page describing the UniProt DAS server and listing the data sources it provides. (If you look at the source of the page, you will see that the browser view is actually an XSLT transformation of the DAS XML returned by the dsn command. The use of this technology with DAS servers is becoming increasingly common.)
You are now presented with a very simple XML file that contains the sequence of the protein Q14974 (Importin beta-1 subunit). This is the primary function of the UniProt DAS server. Note that the server also indicates the length of the protein as an integer and provides a version in the form of an MD5 digest of the sequence. This is used to allow DAS clients to allow a check that all the DAS sources are referring to the same version of the protein sequence as the reference server.
In addition to being a reference server, the UniProt DAS server also acts as an annotation server, allowing the client to query details of features in the UniProt knowledge bank. This also incorporates InterPro features.
Take a careful look at the result of this search. You should be able to find positional features (note that the start and end coordinates of each feature are included) including Swiss-Prot annotation, non-positional features such as a description of the protein (with start and end coordinates of 0) and also literature references relating to the protein.
It is possible for a DAS client to query a coordinate range on an entry point.
Take a look at the definition of the features command in the DAS 1.53 specification. (Keep this page open - you will find this useful later). Use this information to modify the previous query to include only the features located between residues 20 and 100 on the protein sequence.
Using the information from the DAS 1.53 specification, modify your query to retrieve the positions of alpha helices (feature ontology term 'SO:0001117') across the entire Q14974 sequence.
As indicated above, the purpose of DAS is to allow you to retrieve protein information from multiple sources at the same time. An example of a completely separate but compatible DAS annotation server that is able to contribute further annotation of the same protein is given in the next task:
Here you will find additional annotation of the same protein from the MSD Motif database at the EBI. Notice that in this case the version is not the same as the version from the UniProt DAS server for the same protein accession. This could mean that the MSD DAS server is not providing annotation on the same version of the sequence being served from the UniProt DAS server. A typical DAS client should note this discrepancy and warn the user that the protein sequence being annotated is possibly not the same.
This is the annotation server based at UCL in London that also provides features for the same protein. In this case (at the time of writing) the version is the same as the UniProt DAS server.
Clearly there are potentially many different DAS servers that may provide annotation that is useful for you. The next section describes how you can manually find DAS services.
Finding DAS sources – the DAS Registry Service at the Sanger Institute
“The purpose of the DAS registration service is to keep track which DAS services are around, which DAS commands they are understanding and about the coordinate systems of their data.”
The DAS Registration Server allows DAS clients to discover DAS services. At the same time it provides an elegant browsable web interface that allows you to search for DAS sources, test their status, examine their reliability and learn more about what they offer. At the time of writing, a total of 383 servers from 47 institutions in 18 different countries are included in the registry.
The following activities will familiarize you with the content of the registry, both the 'human' interface and also the web-service interface that can be queried by DAS clients (and possibly your client code in the exercises this afternoon).
First of all, a few exercises to familiarize you with the DAS Registry web site:
Navigate to The DAS Registration Server
Follow the "list" ... "list sources" menu item at the top of this page. On the resulting page, find all DAS annotation servers that use UniProt as their 'authority'.
Select any one of these services and follow the information link ('i' in a blue circle) to find out details of this server including how reliable it has been recently.
For the same DAS service, use the DAS Registration Server to test its current capabilities to find out if it is supporting all of the functionality that it claims to at the present time. (Start with the green tick).
The DAS Registration Server allows DAS sources to be organised into projects, which can also be used at the client level to request annotation from logically associated DAS sources, such as DAS sources that are associated with the BioSapiens Project.
To view all the projects currently described in the DAS Registration Server, take a look at: http://www.dasregistry.org/listProjects.jsp
Of course different DAS data sources serve sequences and features based upon numerous different accession and coordinate systems. The DAS Registration Server provides functionality to keep track of this and to ensure a common nomenclature.
Look at the response to the DAS Registry coordsys command: http://www.dasregistry.org/coordsys/CS_DS6. On this page you are presented with a description of a specific accession system, namely the UniProt coordinate system in this case. You will also find a link that allows you to view a list of all the DAS servers that implement this coordinate / accession system.
The next few exercises will allow you to examine some of the web service capabilities of the DAS Registration Server.
Note that the DAS Registration Service provides both a simple DAS style web service allowing you build a simple URL request and receive an XML response, as well as a SOAP (Simple Object Access Protocol) web service. Here we will focus on the DAS style web service.
For the DAS style web service functionality provided by the DAS Registration Service, XSLT style sheets are used to transform the XML response into an attractive web page for viewing in an internet browser. For all of the following activities, please look at both the view in the web browser as well as the source of the page, which will reveal the XML response that can be parsed in suitable client code.
Look at the response to the DAS Registry sources command: http://www.dasregistry.org/das1/sources. As indicated above, the view you are seeing in the browser is a transformation of the XML response into an attractive HTML view. Right click on the browser window and click on 'View Source'. (The exact details of this depend upon the internet browser you are using).
You should see that the XML response provides a <SOURCES> XML file, containing any number of <SOURCE> elements. Each <SOURCE> element describes a single data source, providing parsable information including:
- The source URI as defined in the registry, of the form DS_NNN;
- The title and a description of the data source;
- Details of the datasource maintainer;
- The coordinate system used by the data source (note that this is defined in terms of DAS Registry registered coordinate systems, e.g. "UniProt" has been assigned the URI CS_DS6);
- The capabilities (implemented commands) of the data source;
- A range of additional properties for the service.
You can also perform the sources query for a single datasource: http://www.dasregistry.org/das1/sources/DS_409
For details of web services and WSDL:
Finally, you may like to take a look at the DAS Registry Server Help with Scripting Page that provides more information about how to use the available web services in your own scripts.
Philip Jones, EMBL-EBI, Hinxton, UK. February 2008