[BioC] [limma] heatDiagram function doubts

Marcelo Luiz de Laia mlaia at fcav.unesp.br
Tue Feb 3 20:17:14 MET 2004

I run a heatDiagram and I am with some doubts.

First, when I adjust primary=1 in the function heatDiagram, some of the shown genes are different from those shown by the function topTable when I adjust the coef=1. The opposite also happens. (I considered the same amount of genes in both cases). Why?

Second, when I adjust primary=2 and I execute the function heatDiagram, the genes of the other columns cannot be significant. Correct?

Please, find below the script that I used and information on the design of the experiment, case this is necessary to help me.

Experiment Design

          1day 2day 3day

          Rep1 Rep1 Rep1
UnTreated Rep2 Rep2 Rep2
          Rep3 Rep3 Rep3

          Rep1 Rep1 Rep1
  Treated Rep2 Rep2 Rep2
          Rep3 Rep3 Rep3
2 treatment (treated and untreated); 3 repetitions, and 3 times.
My script (step-by-step)

>files <- dir(pattern="\.tol$")
>RGtol <- read.maimages(files, columns=list
>printer <- list(ngrid.r=4, ngrid.c=5, nspot.r=16, nspot.c=24,
+ndups=2, spacing=1, npins=20, start="topleft")
>layout <- list(ngrid.r=4, ngrid.c=5, nspot.r=16, nspot.c=24)
>genes.names <- read.table("genes_names.txt", header = TRUE, 
+sep ="\t", colClasses = "character")
>genenames <- uniquegenelist(genes.names[,"Name"],ndups=2,spacing=1)
>MAtol <- normalizeWithinArrays(RGtol, layout)
>MAtol2 <- normalizeBetweenArrays(MAtol,method="scale")
>design <- model.matrix(~ -1+factor(c(1,1,1,2,2,2,3,3,3)))
>colnames(design) <- c("time1","time2","time3")
>cor.tol <- dupcor.series(MAtol2$M,design,ndups=2,spacing=1,initial=0.7,
>fit.tol <- lmFit(MAtol2,design, ndups=2,spacing=1, +correlation=cor.tol$cor,weights=NULL,method="ls")
>contrast.matrix <- makeContrasts(time2-time1, time3-time2,time3-time1,
>fit2.tol <- contrasts.fit(fit.tol,contrast.matrix)
>eb.tol <- eBayes(fit2.tol)
>time2.time1tol <- topTable(eb.tol, coef=1, number=50, +genelist=genenames,adjust="fdr")
>time3.time2tol <- topTable(eb.tol, coef=2, number=50, +genelist=genenames,adjust="fdr")
>time3.time1tol <- topTable(eb.tol, coef=3, number=50, +genelist=genenames,adjust="fdr")
>clas.tol <- classifyTests(eb.tol)

Marcelo Luiz de Laia, M.Sc.
Dep. de Tecnologia, Lab. Bioquímica e de Biologia Molecular
Universidade Estadual Paulista - UNESP
Via de Acesso Prof. Paulo Donato Castelane, Km 05
14.884-900 - Jaboticabal, SP, Brazil
PhoneFax: 16 3209-2675/2676/2677 R. 202/208/203 (trab.)
HomePhone: 16 3203 2328 - www.lbm.fcav.unesp.br - mlaia at yahoo.com

More information about the Bioconductor mailing list