Co-expression network using Cytoscape

Question

[EpiExplorer]

Hi

I have mrNA expression and LncRNA expression data from RNASeq. I have calculated the correlation analysis between the differentially expressed genes and the differentially expressed LnCRNAs. I am interested to visualise the correlation matrix in form of co-expression network using Cytoscape.

I would like to know how can I do that? Basically what should my input file look like and what format should it be?

Any help for this protocol will be highly appreciated.

Thanks.

Answer 1

[Wietje]

Your input data can just be in .txt format and you basically just need to include the mRNA and lncRNA (so one identifier in each column) and in a third column the correlation value. That's enough to build a simple network in Cytoscape. Just import the file with FILE -> IMPORT -> NETWORK -> FILE.

You can then add additional information about your network with FILE -> IMPORT -> TABLE -> FILE, for instance if you want to change the colour or shape of your nodes depending on different features, e.g. are they a lncRNA or a mRNA.

I hope this was helpful to you.

[EpiExplorer]

Hi Wietje,

Thanks for the answer. I am new to Cytoscape so my next question may sound stupid. The text file I have have the LncRNAs in first column and the other columns consist of header name as gene names and value in those columns the values as correlation values. Now since the file I have is kind of matrix of correlation of LncRNA vs genes, how can I represent this in the form of network. I am sure there should be way to tell cytoscape what columns is what. Could you please help with this?

Thanks for your time.

[Wietje]

when you upload network information you can chose what the different columns represent and how they are separated and if there is a header or not. Cytoscape offers very detailed and helpful tutorials to get you started - you might want to considering going through this one.

I am not quite sure if I understood correctly what your data looks like - can you post just a few lines to give me a better idea? If it's in matrix structure you should reorganise it into a long format:

gene1 | lnc1 | correlation

gene1 | lnc2 | correlation

gene2 | lnc1 | correlation ...

[EpiExplorer]

[enter link description here][1] First column has LncRNAs IDs, next 5 columns have correlation values between each LncRNA across 5 different genes in the 5 columns.

Image link is below:![enter image description here][2]

Sorry I am unable to send the link for some reason.

[Wietje]

It's like I said, you would need to change from this matrix structure (wide format) to long format, which can easily be done with R. Stackoverflow and the R-cookbook both have helpful posts on how to do this. If the answer on top is a solution to your initial problem, please consider upvoting it.

[Ram]

Please do not paste screenshots of tabular data. See this guide on how to paste tabular data