[BioC] Computing a diference of 2 differences from a CpG Island experiment using LimmaGUI

[Darlene Goldstein]

hi.......

Richard Friedman <friedman at ...> writes:

> 
> Dear Bioconductor Users,
> I am trying to use LimmaGUI to analyze a CpG island experiment.
> 
> I wish to calculate the statistical significance associated with the 
> comparison
> (Tumor Undigested-Tumor Digested) vs. (Normal Undigested-Normal 
> digested).
> (Undigested has both methylated and unmethylated CpG islands. Digested 
> has only
> Unmethylated CpG islands. So Undigested-Digested gives the extent of 
> methylation.
> 
> Comparing tumor to normal gives the change in methylation associated 
> with cancer.
> 
>  
> 
> Mytarget matrix is:
> SlideNumber FileName Cy3 Cy5 Date
> 
> 18370n_13430283.spot Ndigest Nundigest  10/3/06
> 28370n_13430572.spot Nundigest  Ndigest 10/3/06
> 38370t_13469544.spot  Tdigest Tundigest10/17/06
> 48370t_13469546.spot  TundigestTdigest 10/17/06
> 
>  

<snip>

You are trying to estimate an interaction (ie, 'difference of differences'). The
 message explains why it is not possible to do this with the present design: 
the design is unconnected.  The present design looks like:

ND ---> NU
ND <--- NU

TD ---> TU
TD <--- TU

so there is no path connecting any N with any T.  To estimate the interaction
you will also need hybs like Nx ---> Ty (and/or vice versa), where x and y are D
and/or U.  Then you can use limma/limmaGUI to estimate the difference you are
interested in.

Best regards,

Darlene

--
Darlene Goldstein
École Polytechnique Fédérale de Lausanne (EPFL)
Institut de mathématiques
Bâtiment MA, Station 8        Tel: +41 21 693 5565 
CH-1015 Lausanne              Fax: +41 21 693 4303
SWITZERLAND