Hi,
I've analyzed for an investigator of my institution some expression data for DGE with limma. Afterwards, I've used limma mroast function to perform GO enrichment analysis in the same expressionSet and design.
I've found some interesting GO terms, but the investigator whose arrays I'm analyzing would like to have the results in a more graphical way, that takes into account what he calls the strength of the GO term (that I think is something like a p-value, but maybe some kind of mix between p-value and number of genes for this GO term).
I could represent these GO terms as circles with color representing this "strength", but I don't know how to connect the different circles taking into account the structure of GO.
I've read about some packages, as clusterProfiler, but I would like to use the data generated by mroast, and I've not been able to convert my mroast results into the data needed for representing them with clusterProfile.
Is there any package/function to graphically represent the data obtained with mroast?
Thank you in advance
Best regards ,
Elena
Dear Elena,
from my personal experience (i think it is more practical) and i have used the mroast testing to find KEGG pathways with differential expression, and then use the package pathview to plot a specific KEGG pathway of your interest and color the nodes by their fitted expression estimates, but i dont know if that interests you. Regarding the graphical representation of GOterms, i would suggest you the RamiGO R package (http://www.bioconductor.org/packages/release/bioc/html/RamiGO.html)- but i dont know specifically how it can handle the output from mroast, because as far as i know it uses a list of "over-represented" terms from which construsts a graph from them and their parent terms.
On the other hand, you could use another packages for overepresentation of GOterms, such as the GOstats and topGO (which iclude graphical representation via various other packages, such as Rgraphviz package).