[BioC] Bioconductor

James MacDonald jmacdon at med.umich.edu
Wed Nov 12 21:47:43 MET 2003

There are a couple of things you can do. Using the siggenes package you
can take your log ratios and calculate one-sample t-statistics, where
you are testing to see if the mean of the ratios differs from zero. This
is probably the easiest thing to do.

If you want to convert your data back to log intensity values, you can
then either use siggenes or multtest. In this case you would probably
want to use a two-sample paired t-test because there is a clear
dependence between the two samples hybridized to a given chip.

It is simple to convert your M and A values back to log intensity
values. A function to do so would look something like this:

convert.back <- function(M, A){
        G <- NULL
        R <- NULL
        for (i in 1:length(M)){
                G[i] <- (2*A-M)/2
                R[i] <- (2*A + M)/2



James W. MacDonald
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
7410 CCGC
Ann Arbor MI 48109

>>> Robert Cribbie <cribbie at criblab.yorku.ca> 11/12/03 02:13PM >>>

I am a "newcomer" to the field of microarray analysis and I have run
a snag trying to analyze some cdna data.

To determine if the same gene in two different groups is
expressed I assume that I am comparing the expression levels of the
groups after normalization. However, when I use bioconductor to
the data (e.g., maNorm) I only end up with matrices of log ratios
(maM). How do I compare the expression levels of the two groups if I
have the log ratios (and not the normalized expression levels).

Thanks for any help that you can provide.


                      Robert A. Cribbie

Assistant Professor		  Associate Coordinator
Department of Psychology	  Statistical Consulting Service
York University                   Institute for Social Research
Toronto, Ontario, Canada          York University
M3J 1P3

Phone: 416-736-5115 (x88615)      E-Mail: cribbie at yorku.ca 
Fax: 416-736-5814                 Website:

Bioconductor mailing list
Bioconductor at stat.math.ethz.ch 

More information about the Bioconductor mailing list