[BioC] limma vs. maanova for expirement with both contrasts and continuous variables

Gordon Smyth smyth at wehi.EDU.AU
Thu Jan 25 06:22:57 CET 2007


Dear Mark,

There's nothing in the limma User's Guide about continuous covariates 
because they don't arise that often and because it's so easy that 
there isn't much to say. You just include the continous covariate in 
the design matrix directly.

E.g., if you already have a design matrix you could add the BAC column using

   designnew <- cbind(design,BAC=BAC)

If you're using model.matrix() to construct the design matrix, you 
just include BAC as a term as in

   design <- model.matrix(~Line+Treatment+BAC)

(Here Line and Treatment are factors, but BAC is not.) etc etc.

Then the model fit will include a column for BAC. It's all pretty 
much equivalent to so-called "analysis of covariance".

Best wishes
Gordon


>[BioC] limma vs. maanova for expirement with both contrasts and 
>continuous variables
>Kimpel, Mark William mkimpel at iupui.edu
>Wed Jan 24 19:31:18 CET 2007
>
>I have been asked to analyzed a dataset with the following design:
>
>3 rat lines (P, NP, W)
>2 treatments (E, H2O)
>2 time points (1, 2)
>2 technical replicates for each data point
>
>Also, we have the ability to measure the actual blood alcohol content
>(BAC) at each time point.
>
>Because we will be sacrificing rats at each time point, this is not a
>true repeated measures experiment as different rats will be used for
>time point 1 than for time point 2.
>
>The response variable, protein expression, is continuous.
>
>This is a much more complicated design than I am used to working with.
>Which BioC package would be the easiest to use in analyzing this data? I
>am used to LIMMA, but I cannot find anything in the LIMMA vignette that
>describes incorporating the continuous variable of BAC into a design.
>Would MAANOVA work? Or, is there something better?
>
>Thanks,
>Mark
>
>
>Mark W. Kimpel MD
>
>Official Business Address:
>Department of Psychiatry
>Indiana University School of Medicine
>PR M116
>Institute of Psychiatric Research
>791 Union Drive
>Indianapolis, IN 46202
>
>Preferred Mailing Address:
>15032 Hunter Court
>Westfield, IN  46074
>(317) 490-5129 Work, & Mobile
>(317) 663-0513 Home (no voice mail please)
>1-(317)-536-2730 FAX



More information about the Bioconductor mailing list