[BioC] normalization in limma

Gordon Smyth smyth at wehi.edu.au
Tue Aug 19 10:15:12 MEST 2003

At 03:05 AM 19/08/2003, cxd55 at cwru.edu wrote:
>Dear all:
>I'm trying to use limma package. But my intensity data is not two-colored, 
>but single color(gray), like the following:
>Gene1   2596.87    2429.06    2436.62  2716.21      2508.85     2391.20 
>       2515.86     2583.28
>Gene2   1083.47    1071.37    1087.75  1022.47      1144.77     1078.31 
>       1007.18     841.17
>How should I do the normalization in Limma? Any guidance is appreciated.

You need to read the data first into a data frame or a matrix, call it 'x'. 
I will assume that you have done this using 'read.table' or 'read.matrix'. 
You also need to make sure that there are no negative or zero intensities. 
If there are, then you should add a small constant to the negative 
intensities to make them positive. Then

e <- log(x,2)
e <- normalizeQuantiles(e)

This implements quantile normalization for the single-channel data and 
produces normalized expression values on the log-2 scale.


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

More information about the Bioconductor mailing list