jmacdon at med.umich.edu
Mon Jul 19 16:23:20 CEST 2004
Yes boxplot(abatch) does only use 5000 of the intensity values. You can
see how this is done by typing getMethods("boxplot") at an R prompt.
I think a random sample of 5000 is sufficient to characterize perhaps
500,000 data points (if not, I guess we Statisticians should pack up our
bag of tricks and go home!). This is sort of an ironic question, given
the fact that most microarray experiments attempt to characterize the
expression of thousands of genes for a potentially vast population,
using maybe a total of 2 or 3 arrays. ;-D
hist(abatch) does use all of the data points. Again, using
getMethods("hist"), you can see that plotDensity.AffyBatch() is used to
make the plot. Typing plotDensity.AffyBatch at an R prompt will show
that there is no subsetting of the data.
James W. MacDonald
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
Ann Arbor MI 48109
>>> "Matthew Hannah" <Hannah at mpimp-golm.mpg.de> 07/19/04 07:46AM >>>
Just a quick question. I just noticed that boxplot(affybatch), returns
for n=5000. Is this correct? If so does this provide a sufficient
of all intensities on the chip, and how are the probesets(or probes?)
Also does hist(affybatch) also use less than all of the data?
Is there a quick method to access summary stats, 5 number + mean, of
I'm using R 2.0 and BioC devel on ATH1 affy chips (22k probesets).
Bioconductor mailing list
Bioconductor at stat.math.ethz.ch
More information about the Bioconductor