[BioC] methods for Differential expression

[Naomi Altman]

I do not know whether there is consensus on the best method.

Limma, SAM and t-tests (for 2 population problems) are popular, as a 
permuations tests such as found in multtest.  SAM and Limma give 
similar results for simple two population and ANOVA problems if 
calibrated similarly, although in my experience SAM is more 
conservative.  t-tests will give different answers because there are 
usually a large number of genes with very small variance, and the 
moderated denominator will render these non-significant.

Personally, I usually use limma with single-channel analysis, because 
it is the most flexible and I think the model is reasonable.  SAM is 
fine for reference designs with only 1 level of replication.   If a 
t-test is appropriate, so are Limma and SAM.

MAANOVA is another good option, and the documentation indicates it 
can handle more complex models than Limma.

--Naomi

At 08:24 AM 7/24/2006, E Motakis, Mathematics wrote:
>Dear all,
>
>I would like to ask which is, at the moment, the most popular method to
>identify differentially expressed genes for two colour cDNA microarrays. Is
>"limma" the method that one would "trust" more in terms of identifying DE
>genes and at the same time obtain a small number of false
>positives/negatives?
>
>Assuming the data are calibrated, do limma (if this is the best method) and
>simple t-test (test of the log transformed intensities) give very different
>results?
>
>Thank you,
>Makis
>
>
>
>
>----------------------
>E Motakis, Mathematics
>E.Motakis at bristol.ac.uk
>
>_______________________________________________
>Bioconductor mailing list
>Bioconductor at stat.math.ethz.ch
>https://stat.ethz.ch/mailman/listinfo/bioconductor
>Search the archives: 
>http://news.gmane.org/gmane.science.biology.informatics.conductor

Naomi S. Altman                                814-865-3791 (voice)
Associate Professor
Dept. of Statistics                              814-863-7114 (fax)
Penn State University                         814-865-1348 (Statistics)
University Park, PA 16802-2111