Creating interaction network using cytoscape

Question

[piyushjo]

Hi,

I have expression data from 65 patient. I am interested in creating protein-protein interaction network based on known interaction (STRINGdb) and expression (RNA-seq). My main purpose is to identify what interactions are present in tumor vs normal cell, as I have expression data for both. I am focusing on respiration pathway genes only. Could anyone guide me what would be a better way using Cytoscape or if not what is a better alternative?

Thanks so much!

[karthic]

Hi,

You can add ClusterOne app to the Cytoscape and do the cluster analysis and plot the interaction map

KK

[piyushjo]

Hi karthic. I am interested in using the microarray data to build PPI or transcriptional inference network. The ClusterOne app uses Mass Spec data.

[karthic]

Hi Piyush,

you can study PPI using ClusterOne. First submit your proteins to Sting DB and fetch the interaction table. Then submit the interaction table to Cytoscape and run ClusterOne on that.

Sample input is below..

node1   interaction node2   combined_score
DappuP317081    combined_score  DappuP313254    0.637
DappuP64475 combined_score  TRY4A   0.634
DappuP56449 combined_score  TRY4A   0.634

Answer 1

[scooter]

Greetings,

Sure here you go:

1. add the stringApp to Cytoscape 3.7.0
2. in the Network panel search bar, select STRING Protein Query
3. paste your genes of interest into the text box. Note that you should probably filter your gene list to those that show significant fold change and p-Value
4. once the network is downloaded from string, import your expression data as a table (File->Import->Table from FIle...) and make sure to map your gene identifiers to the network's "Query Term" column.

Now you can use Cytoscape's Styles to show the fold change, select subnetworks of interest, etc. Also note that the stringApp allows you to do enrichment analysis also.

-- scooter