Common target genes

Question

[Nicolas Rosewick]

Hi,

I've a list of ~50 genes and I want to know their common downstream or upstream target genes (within all known pathways). My idea was to download for every gene, the list of genes that interact with it, and intersect the results. But I don't know where to start (KEGG, REACTOME,...) I know it's possible to do it with IPA but I'm looking for a free solution.

Thanks

Answer 1

[B. Arman Aksoy]

You can accomplish this via the Pathway Commons 2 web service. The query type of interest to you is the commonstream in either upstream or downstream directions. For example, the following gives you the common downstream of EGFR and ERBB2:

http://www.pathwaycommons.org/pc2/graph?source=EGFR,ERBB2&kind=commonstream&direction=downstream&format=BIOPAX

or if you would like to get this data in a simple interaction format, you can simply use this query instead:

http://www.pathwaycommons.org/pc2/graph?source=EGFR,ERBB2&kind=commonstream&direction=downstream&format=BINARY_SIF

or one with more identifiers to the nodes:

http://www.pathwaycommons.org/pc2/graph?source=EGFR,ERBB2&kind=commonstream&direction=downstream&format=EXTENDED_BINARY_SIF

You can either script this via your programming language of choice, or Paxtools library (Java) or PaxtoolsR. If you rather work with a graphical user interface, then I suggest you either use ChiBE 2 or CyPath2 -- both of them will allow you to run these queries easily.

[Nicolas Rosewick]

Thank you that's great. Just a quick question about pathway commons 2. How can I specify to have the gene symbol in place of uniprot id url (for example: http://identifiers.org/uniprot/Q99075 gives me HBEGF).

Thanks

[B. Arman Aksoy]

Right now it is not possible to request another identifier through the web service, but the last example which produces EXTENDED_BINARY_SIF, will help you get this. In that file, in addition to the standard identifiers, you also have access to other identifiers associated with the node (like HGNC, Entrez Gene ID and etc.) It requires a little bit more parsing, but has much more information compared to simple BINARY_SIF.

Answer 2

[Larry_Parnell]

KEGG and REACTOME are indeed good places to begin to identify the genes encoding proteins that reside in pathways just upstream or downstream of your set of 50 genes, with regard to information flow. The word "target" could imply transcription factor (TF) binding, or your gene encodes a TF. If this is what your genes and their analysis entails, you'd need a different approach, but I'm not certain this is your intention.