[BioC] Re: using MSK files with "mas"

[t-kawai at hhc.eisai.co.jp]

Hi, Lana,

There are two points to be cared in this procedure. One is that
probe-masking should be done after the quantile normalization. RMA quantile
normalization does not provide what you want if there are some NA values.
The other point is to specify "na.rm=TRUE" in the step of summarization.
Unfortunately, current version of C-coded rma() function does not support
na.rm option (correct?). Instead of rma() you can use computeExprSet()
command. In summary, the following script will be your help, I hope. Now I
present my masking script in the form of a function.

#
#     mask.R
#
#     x: AffyBatch object
#     mfile: maskfile name
#
mask <- function(x, mfile, skip=2) {
      m <- scan(mfile, skip=skip, list("", ""))

      ids <- array(0, 0)
      for (i in 1:length(m[[1]])) {
            pname <- m[[1]][i]
            plist <- gsub("-", ":", m[[2]][i])
            parray <- eval(parse(text=paste("c(", plist, ")")))

            ids <- c(ids, pmindex(x, pname)[[1]][parray])
            ids <- c(ids, mmindex(x, pname)[[1]][parray])
      }

      intensity(x)[ids,] <- NA

      return(x)
}

##
## RMA with NA values ##
##

abatch0 <- ReadAffy()
abatch1 <- bg.correct.rma(abatch0)
abatch2 <- normalize.AffyBatch.quantiles(abatch1)

abatch3 <- mask(abatch2, "a.msk")

eset <- computeExprSet(abatch3, pmcorrect="pmonly",
      summary.method="medianpolish", summary.param=list(na.rm=TRUE))


### END of script ###


Try this one!


> Hi,
> I have tried masking particular probepairs with  mas5, expresso,and
> computeExprSet, but find in all cases that the probesets with the
> masks get "NA" expression values.
> In other words, I am unable to mask probepairs  without masking the
> whole probeset.
> Earlier mail seemed to indicate that the masked  probepairs worked.
> Does anyone think that I am doing something wrong?
>
> With RMA I get some NA's but do get some probeset  expression values.
> Thanks,
> Lana
>

_______________________________________

Takatoshi Kawai, Ph.D.

Senior Scientist, Bioinformatics
Laboratory of Seeds Finding Technology
Eisai Co., Ltd.
5-1-3 Tokodai, Tsukuba-shi,
Ibaraki 300-2635, Japan

TEL: +81-29-847-7192
FAX: +81-29-847-7614
e-mail: t-kawai at hhc.eisai.co.jp