Cytoscape and the STRING database

Question

[hmiller]

I am a new user of Cytoscape and I have searched the STRING Protein query with a list of differentially expressed genes I am interested in creating a network with. There are a number of these genes that don't get identified in STRING, so my network is missing them.

Is my understanding correct that this STRING Protein database only has proteins with known protein interaction partners? Do these have to be direct, experimentally validated interactions to be included in STRING? How can I overcome this limitation?

I found that when I expand the network, these same "missing" proteins can now be identified and mapped. Why is that? Can someone better explain what expanding the network means and what it is used for?

Thanks!

Answer 1

[damian.szk]

Hi,

STRING carries all translated genes, not only the ones that have interactions. If the protein does not have any interactions (at all or to the proteins in your network) it will simply appear disconnected from other nodes in your network.

I found that when I expand the network, these same "missing" proteins can now be identified and mapped.

I'm not sure what is happening here, but it could be that you initially search with identifiers that STRING does not recognize. Otherwise they have to be included in your network (albeit maybe disconnected).

However if the proteins are recognized: In Cytoscape there is an option "singletons". If it's turned off (on by default), proteins without connections to your network won't be shown. The same option on STRING website is called "hide disconnected nodes in the network" and can be found in the settings tab.