[BioC] multi-level design - a simplified question - corrected table

[Gordon K Smyth]

> Date: Wed, 30 Jul 2014 22:01:34 +0000
> From: "Rao,Xiayu" <XRao at mdanderson.org>
> To: "bioconductor at r-project.org" <bioconductor at r-project.org>
> Subject: [BioC] multi-level design - a simplified question - corrected
> 	table
>
> Hello all,
>
> I do need some help on analyzing such unorganized data. Please help me out. Thank you so much!
> I basically followed the analysis of multi-level experiments in limma user guide. But I do not feel right about the code below. Please give me some suggestions.
>
> # I want to compare Normal vs. Tumor negative,  and Normal vs Tumor positive. There are partial pairing (subject) and batch effect (chip).
> Treat <- factor(paste(targets$chip,targets$type,sep="."))
> design <- model.matrix(~0+Treat)

No, this isn't correct.  If you need to correct for a batch effect (and 
have you checked that you really need this?), then it should be

   design <- model.matrix(+0+type+chip)

where type and chip are both factors.  Then, when you take contrasts later 
on, you simply compare the type levels that are relevant.

Or better still,

   type <- relevel(type, ref="N")
   design <- model.matrix(~type+chip)
   corfit <- duplicateCorrelation(y,design,block=targets$subject)
   fit <- lmFit(y,design,block=targets$subject,correlation=corfit$consensus)
   fit <- eBayes(fit)
   topTable(fit, coef="typeTneg")
   topTable(fit, coef="typeTpos")

Best wishes
Gordon

> colnames(design) <- levels(Treat)
>
> corfit <- duplicateCorrelation(y,design,block=targets$subject)
> corfit$consensus
> fit <- lmFit(y,design,block=targets$subject,correlation=corfit$consensus)
> cm <- makeContrasts(TposvsN=(a1.Tpos+a2.Tpos+a3.Tpos)/3-(a1.N+a2.N)/2, TnegvsN=(a1.Tneg+a3.Tneg)/2-(a1.N+a2.N)/2, levels=design)        ????
> fit2 <- contrasts.fit(fit, cm)
> fit2 <- eBayes(fit2)
> topTable(fit2, coef=1, sort.by="p")
>
> sample	type	subject	chip
> s1	Tneg	1	a1
> s2	N	1	a1
> s3	Tpos	2	a1
> s4	N	2	a1
> s5	Tneg	3	a1
> s6	N	3	a1
> s7	Tpos	4	a1
> s8	N	4	a1
> s9	Tpos	5	a2
> s10	N	5	a2
> s11	N	6	a2
> s12	Tpos	7	a2
> s13	N	7	a2
> s14	Tpos	8	a2
> s15	N	8	a2
> s16	Tneg	9	a3
> s17	Tneg	10	a3
> s18	Tneg	11	a3
> s19	Tpos	6	a3
> s20	Tpos	12	a3
> s21	Tneg	13	a3
> s22	Tpos	14	a3
>
>
> Thanks,
> Xiayu

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}