[BioC] F values from topTable with coef=NULL testing all zero or all equal?
James W. MacDonald
jmacdon at med.umich.edu
Tue Oct 25 15:11:36 CEST 2011
On 10/24/2011 3:21 PM, Michael Muratet wrote:
> This has probably been answered before, but I can't narrow the search
> down enough to find it in the archives. I am looking at the output
> from and documentation for eBayes and realized that the null
> hypothesis for the F statistic is that all the expression values are
> zero. I want to test the hypothesis that all the expression values are
> equal, i.e., rejection implies differential expression. I see from the
> topTable documentation that calling it with coef=NULL will yield an F
> statistic calculation, and I infer from the docs it will include all
> the conditions. What is the null hypothesis: that all the values are
> equal, or that all the values are zero?
I think you misunderstood the help for eBayes(). It states that the
F-statistic tests that all _contrasts_ are zero, which is not the same
as all expression values equal to zero (which wouldn't make sense).
If a contrast is equal to zero, it implies no differential expression
between the samples compared in that contrast (or more correctly, a
sufficiently small F-statistic implies that you are unable to reject the
null hypothesis of a difference in any contrast). Note that you are only
testing those contrasts that you specified via a contrasts matrix, or
specified indirectly via your design matrix.
> Michael Muratet, Ph.D.
> Senior Scientist
> HudsonAlpha Institute for Biotechnology
> mmuratet at hudsonalpha.org
> (256) 327-0473 (p)
> (256) 327-0966 (f)
> Room 4005
> 601 Genome Way
> Huntsville, Alabama 35806
> Bioconductor mailing list
> Bioconductor at r-project.org
> Search the archives:
James W. MacDonald, M.S.
University of Michigan
Department of Human Genetics
1241 E. Catherine St.
Ann Arbor MI 48109-5618
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues
More information about the Bioconductor