How to extract protein interactions from a reactome biopax file?

Question

[lovetl2002]

Hello,

I'm currently working with Reactome database. I want to get the protein interaction pairs of mouse. However, the file Reactome provided for download is not up to date. A lot of the pairs showed in the browser are missing. So I decide to extract these pairs from the BioPax file they provided.

So the question comes. I can import the file to cytoscape and export it to sif format. For example:

http://www.reactome.org/PathwayBrowser/#SPECIES=48892&DIAGRAM=5957113&ID=5957111&PATH=5956939&DTAB=DT

However, in the exported file, the name of the node is protein name. What I want is uniprot ID. So is there a way to export pairs of uniprot ids rather than protein names in cytoscape? Or is there any ease-to-use tool instead of cytoscape to do this?

[787685856]

change toSif to toSIF

Answer 1

[ozgunbabur]

If you are familiar with Java and maven environment, you can try using the biopax-pattern software to convert the reactome BioPAX (level 3) file into SIF with uniprot IDs. It is in the link below.

http://code.google.com/p/biopax-pattern

Once you have the library running, you can convert the owl with the following code:

String dir = "/dir/to/owl/";
SimpleIOHandler h = new SimpleIOHandler();
Model model = h.convertFromOWL(new FileInputStream(dir + "sample-reactome-file.owl"));

CommonIDFetcher idFetcher = new CommonIDFetcher();
idFetcher.setUseUniprotIDs(true);
SIFSearcher s = new SIFSearcher(idFetcher, SIFEnum.IN_COMPLEX_WITH);
s.searchSIF(model, new FileOutputStream(dir + "output.sif"), true);

Here, instead of SIFEnum.IN_COMPLEX_WITH, you can say SIFEnum.values() and generate all kinds of SIF interactions.

[lovetl2002]

Thank you. The problem is that I'm not familiar with Java. I'm familiar with R and Python though.

Answer 2

[rodche]

Hi,

Try the following cool java tool for BioPAX, Paxtools, which is also partially based on Ozgun's biopax-pattern module and can be run from console (terminal).

First, go grab the latest, just updated, (it's paxtools-4.3.1-beta-no-jena.jar) executable paxtools JAR from the BioPAX project, paxtools downloads.

Next, use either toSif (Simple Interaction Format) or toSifnx command (generates a text tab-separated data file, where a single blank line separates two sections: SIF interaction edges and nodes), e.g., toSif:

$JAVA_HOME/bin/java -Xmx2g -jar paxtools.jar toSif 5957113.owl 5957113.sif.txt uniprot

Note that you'd use uniprot as the last parameter, because there are no HGNC Symbol xrefs in that particular data file.

PS:

Run java paxtools.jar without arguments to see all available command; though, the Paxtools library allows much more if you can code in Java.

[O.rka]

Exception in thread "main" 
    java.lang.IllegalArgumentException: No enum constant org.biopax.paxtools.Main.Command.toSif