[BioC] limma topTable problem

[Jenny Drnevich]

Hi Jianping,

I have just been learning how to use limma, and I almost stumbled into your 
problem. Because of the way your design matrix is laid out, you need to 
specify a contrast matrix before running the eBayes function. See page 31 
in the Limma manual for what to do.

Cheers,
Jenny


At 03:07 PM 9/28/2005, you wrote:
>Dear BC list:
>
>I tried limma's lmFit and eBayes with my data set which consists 6 RG_U34A
>chips, 3 of them as condition 1 and the other 3 as condition 2. What I did
>was as the following:
>
> > data <- ReadAffy()
> > eset <- gcrma(data)
>Computing affinities.Done.
>Adjusting for optical effect......Done.
>Adjusting for non-specific binding......Done.
>Normalizing
>Calculating Expression
> > design <- model.matrix(~ -1+factor(c(1,1,1,2,2,2)))
> > colnames(design) <- c("Diff","Prolif")
> > fit <- lmFit(eset,design)
> > fit2 <- eBayes(fit)
> > topTable(fit2,adjust="fdr")
>                    ID        M        A        t      P.Value        B
>1742 J01435cds#4_s_at 15.69638 15.78457 374.0793 1.514150e-12 22.64603
>8022   X14671cds_s_at 14.58179 14.56121 356.7313 1.514150e-12 22.58357
>3700   rc_AA849038_at 14.58663 14.47957 344.0524 1.514150e-12 22.53258
>5919   rc_AI237836_at 15.01864 14.98568 338.1369 1.514150e-12 22.50706
>8209   X57529cds_s_at 14.96041 14.89656 330.1616 1.514150e-12 22.47075
>3655 rc_AA818069_f_at 15.04516 15.02714 325.2206 1.514150e-12 22.44709
>8219    X58200mRNA_at 14.45639 14.41961 322.4039 1.514150e-12 22.43318
>8281     X62145cds_at 13.81953 13.83622 322.0119 1.514150e-12 22.43121
>7941      X06423_g_at 15.00450 14.93952 321.4805 1.514150e-12 22.42855
>8225  X58465mRNA_g_at 13.19960 13.18034 320.3176 1.514150e-12 22.42267
>
>## I liked to see how many genes with adjusted p-value < 0.05
>
> > dim(fit2)
>[1] 8799    2
> > table <- topTable(fit2, number=8799, adjust.method="fdr")
> > sum(table[,5]<0.05)
>[1] 8788
>
>It was unbelievable that 8788 out of 8799 genes are significantly
>expressed. Therefore I tried to look at p-values before eBayes:
>
> > topTable(fit,adjust.method="fdr")
>Error in array(x, c(length(x), 1), if (!is.null(names(x))) list(names(x),
>:
>         attempt to set an attribute on NULL
>
>I am not familiar with lmFit and eBayes and don't know if it is appropriate
>to make a gene list based on adjusted p-values. Is it B value better?
>
>I'll appreciate anyone who can help me out!
>
>Jianping
>
>xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>x Jianping Jin Ph.D.            x
>x Bioinformatics scientist      x
>x Center for bioinformatics     x
>x 3133 Bioinformatics Building  x
>x CB# 7104                      x
>x University of North Carolina  x
>x Chapel Hill, NC 27599         x
>x Tel: (919)843-6105            x
>x Fax: (919)843-3103            x
>x E-mail: jjin at email.unc.edu    x
>xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
>_______________________________________________
>Bioconductor mailing list
>Bioconductor at stat.math.ethz.ch
>https://stat.ethz.ch/mailman/listinfo/bioconductor

Jenny Drnevich, Ph.D.

Functional Genomics Bioinformatics Specialist
W.M. Keck Center for Comparative and Functional Genomics
Roy J. Carver Biotechnology Center
University of Illinois, Urbana-Champaign

330 ERML
1201 W. Gregory Dr.
Urbana, IL 61801
USA

ph: 217-244-7355
fax: 217-265-5066
e-mail: drnevich at uiuc.edu