[BiO BB] GSEA analysis, phenotype labels

[wenhuo hu]

Hi all,

I have a question about GSEA analysis. There is an parameter about
phenotype label (Java version GSEA), it is about the .cls data, with an
extension, e.g. "WT~KO", or "KO~"WT". I tried both way, the results are
quite different. The way described in the original paper on how GSEA works
does not seem give an excuse how this parameter affect the results, and why
this happened. Please help!

Wenhuo Hu
mskcc