[BioC] Interpretation Fold Change - Limma

Christian De Santis christian.desantis at stir.ac.uk
Tue Nov 12 21:48:46 CET 2013


Hi All,

a question about the interpretation of the logFC in limma. I have a microarray experiment with a common reference (Cy5) design.

I have generated the two matrices in the following manner

design.contrast <- modelMatrix(targets, ref="REF")
contrast.matrix <- makeContrasts (BPCa0-BPCa20, BPCa0-BPCa40, BPCa0-BPCa60,  levels=design.contrast)

and fitted the linear model. No problems there.

The contrast matrix looks like the following:
> contrast.matrix [1:4,1:3]
        Contrasts
Levels   BPCa0 - BPCa20 BPCa0 - BPCa40 BPCa0 - BPCa60
  BPCa0               1              1              1
  BPCa20             -1              0              0
  BPCa40              0             -1              0
  BPCa60              0              0             -1

And the fitted coefficients for those contrasts:
> fit.contrast$coefficients[1:3,1:3]
      Contrasts
       BPCa0 - BPCa20 BPCa0 - BPCa40 BPCa0 - BPCa60
  [1,]    -0.07923765    -0.06255957    -0.14165962
  [2,]    -0.18658108     0.10286377    -0.06617045
  [3,]    -0.02428776     0.04029805     0.09705064

For no reasons, i assumed that the output value would represent the log expression of the first group vs the second as in log(BPCa0/BPCa20) and therefore a negative value would mean downregulation in the first group compared with the second. I have evidence, however, to believe that this might not be the case (by checking the M values)...

Could someone please clarify how do i correctly interpret this fold change as it is not straight forward for me to extract this information from the matrix.

Thanks in advance,
Christian


-- 
The University of Stirling has been ranked in the top 12 of UK universities for graduate employment*.
94% of our 2012 graduates were in work and/or further study within six months of graduation.
*The Telegraph
The University of Stirling is a charity registered in Scotland, number SC 011159.



More information about the Bioconductor mailing list