Circos for GSEA

[A-A-ron]

Hi,

I quite like the look of Circos plots and was wondering if the following is possible before I embark on what appears to be a steep learning curve to get the job done.

I have GSEA results from two distinct experiments (same organism, different models).
In place of performing a simple intersection of the results (and displaying a list of gene sets enriched in both datasets) I would like to present a circos plot that displays the underlying similarity in the gene sets at the gene level.

I envisage the plot would be something like this:
-2 hemispheres which represent each experiment.
-The outer track contains the gene set names (color coded for enrichment score or possibly arbitrarily colored but matched between hemispheres)
-The inner track contains the genes within the gene set (color coded for fold-change)
-Connections are made between the hemispheres only if the gene is dysregulated in the same direction

Is such at task possible. Is circos the right tool for the job. As mentioned above the idea is to go beyond GSEA and provide visual information about the similarity of a given gene set between the two models.

Cheers,
Aaron

[A-A-ron]

There have been quite a few views but no replies as of yet.

Is there any further information or clarification required to get some feedback on this?

Cheers,
Aaron

[neavemj]

Hi Aaron,

I think circos is definitely the right tool for the job because of its flexibility - you can basically plot whatever you want. What you have mentioned would require a heatmap track and a link track. Both of these are easily achieved with Circos.

The tricky part would be getting your data into the right format. I'd say a few python scripts to turn your data into a nice table would do the job!

Best,

Matt.