DASMI
max planck institut
informatik
mpii logo Minerva of the Max Planck Society
 

DASMI Frequently Asked Questions

General and usability questions:

Known problems:

If your question is not answered here, please contact us.

 


What is DASMI?[top]

DASMI is short for Distributed Annotation System for Molecular Interactions, a new approch for sharing, integrating, and assessing molecular interaction data. DASMI is based on the Distributed Annotation System (DAS), a client-server system that aims at replacing central data repositories with distributed data sources.
The general idea of DAS and thus DASMI is that data should be kept with their original providers instead of being unified into central data warehouses. While having the benefit of being fast and allowing complex queries, centralized ressources have the drawback of (I) being rather rigid, as new data sources can only be included by the central authority, and (II) demanding a lot of maintenance work, because once the data has been incorporated into the warehouse, it has to be assured that it stays always up-to-date. In a decentralized system like DASMI, new data sources can easily be incorporated, simply be setting up a new server, and the data is always recent, as it is maintained by its original providers. A schematic picture of the DASMI architecture can be seen below.

 


What is DASMIweb?[top]

DASMIweb is a web server for the dynamic retrieval, integration, and annotation of molecular interaction data sets. DASMIweb is a part of the Distributed Annotation System for Molecular Interactions (DASMI). In a nutshell, it allows the user to retrieve interaction data from many distributed ressources, unify these data, and qualitatitively asses them.

 


Which types of molecular interactions are supported by DASMIweb?[top]

At the moment, DASMIweb supports protein-protein and domain-domain interactions. You define the interaction type through the type of your query and your selection of data sources. That is, if you are interested in retrieving interactions for a certain gene or protein, you simply issue a query by one of its database identifiers (e.g. an identifier from UniProtKB, RefSeq, Entrez Gene). If you are interested in interactions for a particular protein domain, you query DASMIweb with a Pfam identifier (PF00001) or name (AAA).

 


Which interaction datasets are included in DASMIweb?[top]

As DASMIweb performs a dynamic integration of interaction data, the answer to this question will change over the time. A list of all currently available resources can be found here.

You can manage datasets that are used for your particular query in the Source Configuration Panel.

 


Are the interactions experimentally observed, curated from the scientific literature, or computationally predicted?[top]

This depends on the individual data source. Some offer data that has been reported by large-scale experimental assays, others provide interactions that have been manually curated from the scientific literature, and yet others provide computational predictions. Please check the data sources page for more information on the individual data sources used by DASMIweb.

 


How can I query DASMIweb?[top]

To make querying DASMIweb as easy as possible, there is only a single query field. You simply insert the identifier of your target protein/domain and click the button. There is no need to specify the identifier system of your query identifier, DASMIweb will do that part for you and automatically try to determine the type of your query. Only if the query identifier is ambiguous, it will ask you to specify it further.

At the moment you can query for interactions using either:

Please also have a look at the respective tutorial page.

 


What is an identifier or coordinate system?[top]

An Identifier system defines the types of the interactors that are participating in interactions. As DASMIweb retrieves the interaction data form various sources, it is crucial that the data provided by each source is properly defined. If, for example, a source has the identifier system "Entrez,Gene_ID", the interactors are genes defined by Entrez Gene identifiers (e.g. 3064). The identifier system "UniProt,Protein Sequence", on the other hand, defines interactions that are reported between UniProtKB proteins (e.g. HD_HUMAN or P43858); interactions between Pfam protein domains are reported in the "Pfam,Protein Sequence" identifier system. An identifier sytem (e.g. Entrez,Gene_ID) is described by the four fields authority, version, type, and organism. Authority is the name of the institution that defines the identifiers of the interactor, for instance, Entrez, UniProt, or Pfam. Type denotes the physical dimension of the interactors, such as "Protein Sequence", "Protein Structure", or "Gene_ID". Version and organism are optional fields to further define the interactors.

Identifiers can be mapped between different identifier systems. For exmaple, genes are the blueprints for proteins, thus an identifier in the "Entrez,Gene_ID" identifier system can be mapped to at least one identifier in the "UniProt,Protein Sequence" identifier system and vice versa. See the mapping entry for further details on the mapping procedure.

 


How does DASMIweb perform the identifier mapping?[top]

If you query DASMIweb with a Uniprot identifier like "HD_HUMAN", it will automatically convert this identifier to a UniProtKB accession number ("P42858"), an Entrez Gene ID ("3064"), a RefSeq ID etc. The advantage if this mapping is that all the servers that only provide interactions in an identifier system different to the query identifier systems will also be incorporated. The disadvantage of this mapping procedure is that it can produce a considerable overhead. For instance, while the mapping from UniProtKB to Entrez Gene is usually one-to-one, mapping the other way round is ambiguous, as one gene may be responsible for several proteins.

A good example is an interaction between two gene interactors represented by their Entrez Gene identifiers, for instance the interaction 1212 3092. The Entrez Gene identifier 1212 can be unambiguously mapped to the UniProtKB identifier P09497. For the gene with the identifier 3092, however, there are six different UniProtKB entries: O00291, A4VCI6, B4DK46, B4DYD7, B4E3I7, and Q8TDA4. If DASMIweb wants to retrieve confidence scores for the interaction 1212 3092 from a server that requires UniProtKB identifiers as input, it will issue six queries to the scoring server, P09497 O00291, P09497 A4VCI6, P09497 B4DK46, P09497 B4DYD7, P09497 B4E3I7, P09497 Q8TDA4 and combine their results.

The basis for the mappings are the iProClass and Pfam databases. The BioThesaurus offers a way to investigate the mappings that are available via iProClass. For instance, this entry shows that the UniProtKB accession numbers P09497, Q6FHW1, the Entrez Gene identifier 1212, and the RefSeq identifiers NP_001825 and NP_009028 all belong to the same entry. Hence, DASMIweb would unify interactions that are reported for either one of these identifiers to one interaction.

 


How does DASMIweb present the resulting interactions?[top]

As this is best explained by pictures, please have a look at the respective tutorial page.

 


How can I manage the data sources that DASMIweb uses?[top]

DASMIweb manages all its data sources in the Source Configuration Panel, which you can open by clicking on the Show Source Configuration button right below the query input field. The Source Configuration lists all sources that DASMIweb knows according to their identifier systems. The tabs in the top of the Source Configuration let you switch between the identifier systems. If you already issued a query to the system, DASMIweb will automatically show you all sources that have the same identifier system as your last query. That is, if your last query was a Pfam ID, the Source Selection will initially show you all sources that support the "Pfam,Uniprot Sequence" identifier system.

Please also have a look at the respective tutorial page.

 


How can I add additional data sources?[top]

In order to integrate your interaction data into DASMIweb, you need make it available as a DAS server. For detailed instructions on how to do that, please read this tutorial (PDF).

Another option of adding your data, if you have it in the form of a valid PSI-MI2.5 XML file, is explained in this tutorial

 


How do I remove existing data sources?[top]

To remove an existing data source, you have to open the Source Configuration and navigate to the tab that contains the identifier system of the source you want to remove. Now you can exclude the results of a specific source by clicking on the checkmark in front of the respective row.

 


How can I apply quality measures to assess the reliability of interactions?[top]

As this is best explained by pictures, please have a look at the respective tutorial page.

If you want to add additional quality measures, also have a look at the next FAQ entry.

 


Can I add additional quality measures into DASMIweb?[top]

Yes. Quality measures are provided in the same way as interaction sources. Therefore, to add additional quality measures, you have to follow the steps listed in this FAQ entry.

 


Why does the drop down menu only list the entry "No confidence scores available"?[top]

Initially the drop down menu only presents the confidence measures that were originally included in the interaction datasets (e.g. the score that a database provider assigned to the interactions in their database). If none of the interaction sources containes these original scores, the drop down menu will be empty.

Please have a look at this tutorial page to see how you can request additional confidence scores.

 


Why do I only see very few coloured squares after I selected a quality measure method?[top]

While there is almost no delay between the initial query and the presentation of the resulting interactions, there is a noticeable time span before all quality scores are available. If, very soon after your query, none or only very few scores are shown after you selected a quality measure, please select the Select confidence measure ... entry again and retry your initial quality measure after a short waiting period.

The reason for the delay is that mapping identifiers between different identifier systems can create a considerable overhead (see this FAQ entry for more details on the mapping procedure). Both currently avaialable confidence scoring servers use the UniProtKB identifier system. Most of the currently available interaction sources, however, provide their data in the Entrez Gene ID identifier system. Therefore, DASMIweb has to map each interaction from the Entrez Gene ID to the UniProt identifier system in order to include the BPscores. In many cases, one Entrez Gene ID maps to more than ten UniProt accession numbers and DASMIweb has to query 100 possible identifier combinations instead of one. If there are many interactions, this procedure may take several seconds and the scores will be available bit by bit.

 


Does DASMIweb provide additional details on individual interactions?[top]

DASMIweb has the potential to show additional details like confidence scores or binding sites on particular interactions. These details, however, are optionally provided by the individual sources and not by DASMIweb itself. Details on an interaction can be requested by clicking on the respective square in the interaction table. A row containing aditional details will be added to the interaction table right below the interaction square. DASMIweb highlights the currently selected interaction by coloring the associated interaction square orange  ^ . The detail area associated with an interaction can be closed by clicking on its orange square.

Please also have a look at this tutorial page.

 


Is there a way to save my personal DASMIweb settings?[top]

Yes, DASMIweb stores your personal configuration, that is, all your sources with their activation states, your last query, the quality measure you selected, additional details you opened, and your myDASMI settings for half an hour. So even if you leave the DASMIweb website in your browser, as soon as you come back within this time frame, everything will be as it was before.

 


Can I export the interactions that DASMIweb retrieved?[top]

Yes, DASMIweb offers you the possibility to export the interactions that you integrated from different sources in various file formats. To export the interactions that are currently presented to you, you have to select you desired export format in the drop down menu Export as ... that is placed on top of the Interaction Area. At the moment, the following export options are available:

 


What is the purpose of the myDASMI rectangle at in the middle of the right border of the DASMIweb screen?[top]

myDASMI is intended as a quickly accessible menu where you can control all important DASMIweb settings. By clicking on the rectangle you can show or hide the quick configuration. myDASMI offers you the following options:

 


Why does DASMIweb only work with activated JavaScript?[top]

DASMIweb heavily relies on a technique called Asynchronous JavaScript and XML (AJAX). This technique allows sending data between the user interface that presents all the interactions and the backend that performs all the computation in an asynchronous manner. This is very important, as it allows DASMIweb to present the results of individual sources as soon as they arrive. This way, even if one server responds immediately and another server takes a rather long time to respond to a query, both results will be presented to you without delay. Unfortunately, there is no way to realize a system like that without JavaScript.

 


If I try to query DASMIweb, it keeps sending me back to the welcome screen after some seconds, even though it was able to determine the type of my query. Is it broken?[top]

This problem can occur if your browser blocks cookies. Please have a look at the next FAQ entry to see why DASMIweb requires cookies and how you can activate them in your brwoser. If the problems persist, please contact us so that we can try to help you and fix potential bugs.

 


Why does DASMIweb require cookies?[top]

In order to establish the asynchronous communication and store your data in sessions, DASMIweb has to be able to uniquely identify users. This identification is achieved by assigning unique session identifiers to each user. These identifiers are stored in a cookie and are subsequently sent to DASMIweb with each request. You do not need to worry about your privacy, however, as the session identifiers are random character strings like B75F74F5D6E33D7DE4BA078E9A719BF5860 and will be deleted after your session has expired.

The steps required to enable cookies depend on your browser.

 


Why do I only see the error message "Permission denied to call method XMLHttpRequest.open" when I try to open DASMIweb?[top]

If you see this message you most likely use either the Mozilla Suite or an old Firefox version (prior to 1.5). These browsers prevent certain JavaScript methods to be executed, causing the respective error. We are working on a workaround to make DASMIweb usable in these browsers.