Almost all FDR values are equal to 1

[Masashi Azuma]

Hello, 

I believe I have successfully ran a GSEA; however, when I look at the results it says that for my wild-type comparison a lot of the gene sets have a q value that is equal to 1. I believe that means all of these are false discoveries? I'm not sure what is causing this as I ran the exact same process on a different time point and it seems to not reproduce this. I have attached a summary below. There is another time point too where this exact problem happens but this time with wild-type and knock-out outputs. 

Thank you so much for the help!

[Anthony Castanza]

Hi Masashi, 

Sorry I missed this. My recommendation would be to only run one component of GO at a time (i.e. MF, or BP, or CC) rather than running them all at once. There is overlap between the sets in the different ontologies that can adversely affect the FDR. There is definitely something odd going on here though, I really wouldn't expect this much of an effect on the FDR. Could you describe your dataset in detail?

-Anthony

Anthony S. Castanza, PhD

Curator, Molecular Signatures Database

Mesirov Lab, Department of Medicine

University of California, San Diego