[R] permax package

Jane Fridlyand janef at stat.berkeley.edu
Wed Mar 13 05:03:03 CET 2002

```Hi, I am having a problem using permax/permcor functions from permax
package.

For the variables that are constant among classes (two classes in my
case),  permutation-based p-value of 0 is outputted.

Artificial Example:

rsp_c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0
,1, 0, 1, 1, 0, 0, 0, 0, 0)

library(permax)
tmp_rbind(rep(1,32), matrix(rnorm(32*1000), ncol=32))
try_permax(data=as.data.frame(tmp), ig1=(1:length(rsp))[rsp==0], nperm=10,
logs=F, ranks=T, min.np=1, ig2=(1:length(rsp))[rsp==1], WHseed=NULL)

> try\$p2[1]
[1] 0

summary(try\$p2[-1])
Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
0.5000  1.0000  1.0000  0.9991  1.0000  1.0000

Is it intended to be that way or is it a bug?

I also wanted to doublecheck:

objects \$p2, \$p.lower and \$p.upper outputted by the permax package
functions correspond to permutation-based p-values where empirical
distribution under null is constructed by permuting labels at random and
taking maximum (minimum) statistic at each permutation -- is it correct? I
seem to be getting fewer significant variables when applying my own (much
slower) functions with the same number of permutations.

Thank you very much

Jane

ps i am using the latest versions of R and of permax

**************************************************************************************
Jane Fridlyand, Postdoctoral Scientist
UCSF Cancer Center, Box 0128 San Francisco, CA 94143-0128
Office: Room N412 Tel: (415)514-3302 Fax: (415)502-3179
**************************************************************************************

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

```