[BioC] RE : Re: question on limma package

[Jenny Drnevich]

Hi Will,

I don't know about B statistics offhand (I never use them), but the 
adjusted p-values are completely dependent on the number of probes 
that you analyze together. In fact, "adjusted" refers to the 
correction for multiple hypothesis testing, which is what you are 
doing when you do a series of linear models on the probes. Check out 
?p.adjust and the references in there for details on different 
adjustment methods. In limma, the unadjusted p-values are only 
slightly affected by the number of other probes analyzed at the same 
time, because it uses a Bayesian correction based on the unexplained 
variation across all probes.

HTH,
Jenny

At 05:27 AM 7/19/2007, William Mifsud wrote:
>Dear Bioconductor,
>
>Further to the previous questions about cutoffs for DE genes,  I was
>wondering to what extent, if at all, do the adjusted (and unadjusted)
>p-values and B statistics depend on the number of probes on the
>array/'present' probes in a particular experiment?
>
>Best wishes,
>
>Will
>--
>
>William Mifsud
>
>_______________________________________________
>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

Jenny Drnevich, Ph.D.

Functional Genomics Bioinformatics Specialist
W.M. Keck Center for Comparative and Functional Genomics
Roy J. Carver Biotechnology Center
University of Illinois, Urbana-Champaign

330 ERML
1201 W. Gregory Dr.
Urbana, IL 61801
USA

ph: 217-244-7355
fax: 217-265-5066
e-mail: drnevich at uiuc.edu