Hi all,

I just installed R 2.5 today and then installed Bioconductor 2.0 using the recommended biocLite(). I then tried to call duplicateCorrelation() from limma, but got this error:

 > corfit <- duplicateCorrelation(MA.nobg.scale,design,ndups=1,block=c(1,1,2,2))
Loading required package: statmod
Error: could not find function "mixedModel2Fit"
In addition: Warning message:
there is no package called 'statmod' in: library(package, lib.loc = lib.loc, character.only = TRUE, logical = TRUE,

It appears that biocLite() didn't get the statmod package from CRAN. Once I downloaded and installed statmod, everything was fine. I don't remember having to do this in the past, so did something get accidentally changed?

Thanks,
Jenny

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

Hi Jenny,

statmod isn't a dependency for limma - it is in the Suggests field, so IIRC, won't be installed automatically.

Best,
Jim

Hi Jenny,

statmod is deliberately a "suggests" rather than a "depends" because only a small number of limma functions use it.

As far I remember, the Bioconductor install behaviour hasn't changed. Maybe I misremember, but I don't think biocLite() has ever installed suggested packages from CRAN. What has changed is that the CRAN installation

install.packages("limma")

will no longer install suggested packages whereas it did in previous versions of R. The R 2.5.0 CHANGE file says:

"The 'Install packages ...' menu item now only installs essential dependencies ('Depends' and 'Imports' but not 'Suggests'), using install.packages(dependencies = NA)."

Best wishes
Gordon