lmFit/mrlm output is vastly easier to work with, when you are dealing
enormous numbers of regressions, in my opinion (having forgotten this
recently and used anova() on a high-dimensional mlm object from base
note that eBayes is independent of lmFit, which is the function that
closely maps to anova(lm(foo ~ bar*baz)) in base R. If you have a
believe that there is a tractable distribution for the variance across
the responses you are fitting, eBayes is an enormous advantage. But
nothing fundamentally to do with the speed and manageability advantage
comes from using lmFit() (or mrlm(), if memory serves -- a robust
1) eBayes is not analogous to anova(lm(y~x)); lmFit() and mrlm() do
2) lmFit() and mrlm() have significant practical advantages over base
All of this is simply IMHO, of course.
On Mon, Aug 22, 2011 at 2:42 AM, carol white <email@example.com>
> Is there any preference for eBayes in Limma package to the anova R
> function? In addition to moderate t-test, is there any other feature
> differentiate these two functions?
> Bioconductor mailing list
> Search the archives:
If people do not believe that mathematics is simple, it is only
do not realize how complicated life is.
John von Neumann<http: www-groups.dcs.st-="" and.ac.uk="" ~history="" biographies="" von_neumann.html="">
[[alternative HTML version deleted]]