Hello,
I am using fdrtool(x, statistic="pvalue") where x is a vector of chisq
p-values.
I can get this command to work with some x, but not with others.
When it does NOT work, I get the following screen
output:> fdrtool(pvals$Chi2.pval, statistic="pvalue")
Step 1... determine cutoff point
Step 2... estimate parameters of null distribution and eta0
Step 3... compute p-values and estimate empirical PDF/CDF
Error in pvt.isoMean(-rawslope, dx) :
NA/NaN/Inf in foreign function call (arg 1)
In addition: Warning message:
Censored sample for null model estimation has only size 0 !
Range of my data:> min(pvals$Chi2.pval)
[1] 0> max(pvals$Chi2.pval)
[1] 0.0498602
I've tried using subsets of pvals$Chi2.pval as x, and fdrtool works
sometimes (the working subsets of values overlap, but fdrtool does not work
for their union). Has anyone else encountered this before? If so, what do I
do about it? Suggestions for alternatives to fdrtool welcome.
There isn't much in terms of R documentation on pvt.isoMean:
fdrtool-internal package:fdrtool R Documentation
Internal fdrtool Functions
Description:
Internal fdrtool functions.
Note:
These are not to be called by the user (or in some cases are just
waiting for proper documentation to be written).
Aliase:
pvt.isoMean, r2t, t2r, pvt.nullfunction, pvt.plotlabels, ecdf.pval,
num.curv, approximate.fit, iqr.fit, get.nullmodel, pvt.fit.nullmodel
Keyword:
internal
Thank you.
ND
--
View this message in context:
http://r.789695.n4.nabble.com/fdrtool-help-tp3161417p3161417.html
Sent from the R help mailing list archive at Nabble.com.
