GSEA Support

[nijibabulu]

Thanks for the outstanding package!

I am trying to formulate an analogous GSEA metric for the output of glmmSeq. Using other methods (edgeR, limma, DESeq2, wilcoxon tests) the test statistic is signed. The statistic is used for ranking the genes for differential expression, and the sign conveniently indicates the direction of enrichment. In the case with using a chi-squared statistic, I am wondering if there can be an analog. So far I have been doing the following for the contrast implied by b and the statistic stat:

sign(res@stats$coef[,b])*res@stats$Chisq[,stat]

Any thoughts on this would be much appreciated!