Frustrated by the poor quality of the output of the desktop version javaGSEA (i.e., low resolution, only in pixel-based format, etc), I have written the function replotGSEA to re-plot data from the javaGSEA desktop version. The function takes three arguments: path, the path to the javaGSEA output folder; gene.set, which is the name of the gene set you want to plot (note: approximate matching is used in a grep-function to determine the appropriate gene set); and class.name, which is the name of the variable to which gene expression has been matched (i.e., the description of the variable that is specified in the .cls file or that is used to pre-rank the data; example: "drug treatment"). An example of the output is below. Please note that R can generate pdfs, allowing for vector-based output (unfortunately I cannot link pdfs in this post and have to use .png format instead). I have posted the code on GitHub. Comments on the output are welcome! I hope you find this useful.
EDIT: I have now updated the tool and included the possibility to use custom metric ranges other than [-1, 1]
EDIT2: Upon feature request, I have now updated the tool and included the possibility to use custom enrichment score ranges
I have a question regarding your script, which is really useful. Is it possible to set the enrichment score y-axis to certain values (e.g. -1 to 1), so that I can create matching graphs for two different pathways? In addition is it possible to change the main title?
Unfortunately, both are not possible in the current version. With regards to the title, I think it makes sense to have the gene set name as main title, and you may export the new plot from R as a pdf, and use for instance Adobe Illustrator to change the title. With regards to the range of the enrichment score plot, I can implement that feature; I will try to do that asap.
Your feature request for having custom ranges for the enrichment score has been implemented in release version 1.4.
Hi, I've tried to use replotGSEA using my results. It's a very nice function but I found and issue I couldn't fix. The heatmap is not centered. I mean, in the original Java plot white is in the middle of the color scale, but in the plot of replotGSEA it is shifted sideways. Does anyone know why and how to solve the problem?
javaGSEA:
ReplotGSEA: