Dear Michael,

I think there is a little flaw in the source code of lfcShrink with DESe2 version 1.22.2. Given one wants to turn off the svalue calculation one would set svalue=F. Still in the code if one sets type="apeglm" the svalue will in any case be turned on without respecting the value of svalue in the lfcShrink function (line 200 of the code):

if (lfcThreshold > 0) {
      message(paste0("computing FSOS 'false sign or small' s-values (T=", 
                     round(lfcThreshold, 3), ")"))
      svalue <- TRUE
      apeT <- log(2) * lfcThreshold

?lfcShrink suggests that one can choose between svalue and p-value.

svalue  
logical, should p-values and adjusted p-values be replaced with s-values when using apeglm or ashr. s-values provide the probability of false signs among the tests with equal or smaller s-value than a given given's s-value. See Stephens (2016) reference on s-values.

Came across this as I wanted to turn off svalues (because it would probably only cause confusion among colleagues and it is not common yet) but no matter if svalue is T or F they will be calculated in both cases. Maybe you can check. Thanks!

It’s not a bug but it’s because you want to perform a threshold test on the posterior value.

We don’t have a way to do that and produce a pvalue