Hi, I am wondering if anyone have used the topconfects package to evaluate the results of a differential gene expression analysis? As I understand it the package calculates confidence intervals around the effect sizes (log fold changes). How do you analyse the results? Such as confect score threshold?

Thanks,

Jon

You can use a threshold, in which case the results are equivalent to using the TREAT method. Any advice for choosing a TREAT threshold also applies to topconfects.

You can also view it as a ranking from most to least important discoveries. For example you could examine the top-most genes in detail, and also use the full ranked list as input to gene-set enrichment.