[BioC] how to design matrix by edgeR

wang peter wng.peter at gmail.com
Tue May 22 20:48:55 CEST 2012


dear Dr.Gordon:
   thank  you for your patient reply, but i still donot understand
what problem in my factor design?i have 3 questions as below:

the 1st question:
   my experiments involves two factors, time and treatment.
   so i define:
 treatment=factor(c(rep('treated',24),rep('control',11)))
 time=factor( c('0h','0h','0h','6h','6h','6h','6h','12h',
 '12h','12h','12h','18h','18h','18h','18h','24h','24h','24h','36h','36h','36h','48h','48h','48h','0h','0h','0h','6h','12h','18h','24h','36h','48h'))

the 2nd question:
i want to get DE cross time considersing control.
and my model.matrix is
design <- model.matrix(~time+treatment)?
or design <- model.matrix(~treatment+time)?

and yours is design <- model.matrix(~time+time:treatment)
which one is correct?


the third question:
      i donot do what is the difference between test by glmLRT
and test by pairedwised test functions?


To test for treatment effect at time 0:

  lrt0 <- glmLRT(y,fit,coef=8)
  topTags(lrt0)

To test for treatment effect at time 1:

  lrt1 <- glmLRT(y,fit,coef=9)
  topTags(lrt1)

thank you very much
if possible, can you give me some detailed material
of glmLRT or glmFIT
now i can only read your paper published recently

shan gao



More information about the Bioconductor mailing list