How to integrate multi omics data in cytoscape

Question

[PK]

Hi all,

I am new to pathway and network analysis. I have differentially expressed genes and mass spec data. i know my bait protein so i want create a network bait as my source and other protein as a target (which i got from my mass spectrometry). on the other hand i knock out the same protein and did the RNA Seq and i got the differentially expressed genes. currently i'm trying to put this two set of data's together. I took all the differentially expressed genes and pasted into genemania i got the matrix but the problem all the genes extending their network. which i don't want. i want to have the genes which have forms any network or not. and if yes is there any of these genes/network forms any interaction with my protein list. Is that make any sense? If you have any question please ask me. Thanks in adavance

[scooter]

Hi there,

If I understand you correctly, you want to see a network which contains all of your mass spec pull-downs (so, essentially a star network with your bait in the middle), then augment that network with any additional proteins that interact with your bait proteins from your differentially expressed list. Is that right? I assume you also want to be able to visualize the expression on the nodes in the resulting network. Here is one approach using Cytoscape desktop that might get you there -- see if this works for you:

1. Create your initial network from your mass spec data. Make sure to add a column in your data set for "Type" or something like that and indicate whether this protein is a bait or a prey (we'll use that later).
2. Paste your DEGs into the stringApp and make sure to set the "number of additional proteins" to 0. I would also set the network type to physical interactions. Now you have two networks that can be merged together.
3. At this point, there are two options. We can merge both networks and filter down to what you want, or filter in the DEG network first. Taking the second approach first, select the gene that encodes your bait protein in the DEG network and extend the selection to first neighbors. Create a subnetwork with just those genes and all edges (only the edges connecting those genes will be included). Now, you can merge that network with your original proteomics network, and (I believe) you get exactly what you wanted.
4. Alternatively, you could merge the two networks first to see if there are any interactions between the proteins you pulled down and your DEGs. This would be a more exploratory step. Then you could filter things down as I described above, or potentially, select all of your original bait and prey proteins (remember that "Type" column?) and extend the selection to first neighbors to create a network with everything that is differentially expressed that interacts with either your bait or prey proteins.
5. To add the expression visualization, once you have your network, just read in your expression table and map the values onto the gene of interest. Note that you may have to be careful with the column in the network you are using for the mapping to ensure that the identifiers you used for your proteomics network and the identifiers you used for your DEG network are compatible.

-- scooter

[PK]

Hi,

I think this what exactly i want. I will try and comeback to you. Thanks for your information and help.