[BioC] Reg: T-statistic using limma

[Claus Mayer]

That Venn Diagram is not very interesting in your case. The intercept will
(unless you centered the data for some reason)  always be significant and is
of little interest. The comparison you are interested is between the groups.
The empty circle in the diagram just means that you have no genes
significant at an Benjamini-Hochberg adjusted p-value below 5%. That can
happen if there are not many differential genes. You should look at the
p-values for the group comparison, eg. look at the toptable for the second
coefficient (that is the log-ratio between treatment and control)...

If you want a picture, the p-value histogram gives you an overview
(hist(fit.eBayes$p.value[,2],nclass=100) or something like that). If that
has a peak at the left, it indicates that some genes are differentially
expresse. If it is flat, it indicates that all your results might as well be
chance.

Probably it makes sense to work through one of the examples giving in the
limma tutorial and make sure you know exactly what you are doing, before you
go on with your own data...

Claus 

> -----Original Message-----
> From: bioconductor-bounces at stat.math.ethz.ch 
> [mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of 
> Rohit Farmer
> Sent: 16 March 2010 15:16
> To: bioconductor at stat.math.ethz.ch
> Subject: Re: [BioC] Reg: T-statistic using limma
> 
> thanks Claus
> 
> i did as u mentioned and got the results ... but when i am running
> 
> results <- decideTests(fit)
> and after that
> 
> vennDiagram(results)
> 
> the vendiagram shows two circles left one shows intercept 
> will the number of all the genes and the right one 
> population.goupsLLO contains zero value and the intersection 
> is also not showing contains zero how can i visualize my results....
> 
> 
> On Tue, Mar 16, 2010 at 7:29 PM, Rohit Farmer 
> <rohit.farmer at gmail.com> wrote:
> > Hi there ... here i am trying to do a moderate t-statistics using 
> > limma i use the following code and got these results but my 
> process in 
> > not able to complete and showing some error at the end
> >
> >> population.goups 
> >> 
> <-factor(c('10LLO1.CEL','11LLO2.CEL','12LLO3.CEL','1Control1.CEL','2C
> >> ontrol2.CEL','3Control3.CEL')
> > + )
> >> population.goups
> > [1] 10LLO1.CEL    11LLO2.CEL    12LLO3.CEL    1Control1.CEL 
> > 2Control2.CEL 3Control3.CEL
> > Levels: 10LLO1.CEL 11LLO2.CEL 12LLO3.CEL 1Control1.CEL 
> 2Control2.CEL 
> > 3Control3.CEL
> >> design <- model.matrix (~population.groups)
> > Error in eval(expr, envir, enclos) : object 'population.groups' not 
> > found
> >> design <- model.matrix (~population.goups) design
> >  (Intercept) population.goups11LLO2.CEL population.goups12LLO3.CEL 
> > population.goups1Control1.CEL
> > 1           1                          0                          0
> >                         0
> > 2           1                          1                          0
> >                         0
> > 3           1                          0                          1
> >                         0
> > 4           1                          0                          0
> >                         1
> > 5           1                          0                          0
> >                         0
> > 6           1                          0                          0
> >                         0
> >  population.goups2Control2.CEL population.goups3Control3.CEL
> > 1                             0                             0
> > 2                             0                             0
> > 3                             0                             0
> > 4                             0                             0
> > 5                             1                             0
> > 6                             0                             1
> > attr(,"assign")
> > [1] 0 1 1 1 1 1
> > attr(,"contrasts")
> > attr(,"contrasts")$population.goups
> > [1] "contr.treatment"
> >
> >> fit <- lmFit(eset, design)
> >> fit.ebayes <- eBayes(fit)
> > Error in ebayes(fit = fit, proportion = proportion, stdev.coef.lim =
> > stdev.coef.lim) :
> >  No residual degrees of freedom in linear model fits
> >
> > 
> ----------------------------------------------------------------------
> > ---
> >
> > Please any help in this regard will be highly appreaciated
> >
> >
> > Rohit
> >
> >
> >
> > --
> > Rohit Farmer
> > M.Tech Bioinformatics
> > Department of Computational Biology and Bioinformatics 
> Jacob School of 
> > Biengineering and Biotechnology Sam Higginbottom Institute of 
> > Agriculture, Technology and Sciences (Formerly known as Allahabad 
> > Agricultural Institute - Deemed University) Allahabad, UP, 
> INDIA - 211 
> > 007 Ph. No. 9839845093, 9415261403 e-Mail 
> rohit.farmer at gmail.com Blog 
> > http://rohitsspace.blogspot.com
> >
> 
> 
> 
> --
> Rohit Farmer
> M.Tech Bioinformatics
> Department of Computational Biology and Bioinformatics Jacob 
> School of Biengineering and Biotechnology Sam Higginbottom 
> Institute of Agriculture, Technology and Sciences (Formerly 
> known as Allahabad Agricultural Institute - Deemed 
> University) Allahabad, UP, INDIA - 211 007 Ph. No. 
> 9839845093, 9415261403 e-Mail rohit.farmer at gmail.com Blog 
> http://rohitsspace.blogspot.com
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: 
> http://news.gmane.org/gmane.science.biology.informatics.conductor
>