[BioC] Problem with gls.series in limma

[Gordon Smyth]

You're using an older version of limma with a newer version of statmod. 
This may have
occured because you have the last Bioconductor release of limma with a more 
recent
version of statmod from CRAN. The problem is that Bioconductor updates its 
packages
only once every 6 months (unless you use the developmental version, which I 
don't
recommend) while CRAN packages are updated as required.

The solution is to update limma and statmod directly from CRAN, which will give
you the current versions of both packages. Just use

  install.packages("limma")

or use the drop down menu if you're using Windows.

BTW, I recommend that you move to mFit() rather than using gls.series() 
directly.

Gordon

At 09:52 AM 23/09/2004, Fangxin Hong wrote:
>Hi there;
>    I tried to use gls.series to get least square fit for each gene, but I
>got the following error message:
>--------------------------------------------------
>Loading required package: statmod
>
>Attaching package 'statmod':
>
>
>         The following object(s) are masked from package:limma :
>
>          matvec vecmat
>
>Error in randomizedBlockFit(y, X, Z, fixed.estimates = FALSE) :
>         unused argument(s) (fixed.estimates ...)
>------------------------------------------------------
>
>Any clue?
>
>Thanks a lot!
>
>
>fangxin
>
>
>--
>Fangxin Hong, Ph.D.
>Bioinformatics Specialist
>Plant Biology Laboratory
>The Salk Institute
>10010 N. Torrey Pines Rd.
>La Jolla, CA 92037
>E-mail: fhong at salk.edu