Hi Diya,
at this stage of your analysis you're leaving the world of
aroma.affymetrix. Please use the Bioconductor mailing list for
questions about limma etc. You have a much better chance to get a
quick and clear answer there.
/Henrik
On Thu, Sep 3, 2009 at 12:22 PM, Diya Vaka<biotechd...@gmail.com> wrote:
>
> Hi,
>
> I have a question.I have 2contral samples(v7,v8) and 2 treatment
> samples(g1,g4).I have the data generated from aroma.affymetrix and I
> have applied log2 on it and load the data for doing ststistical
> analysis by limma.But I dont understand whether to give coef=2or 1 in
> the toptable command.When I give one it gives very little fold change
> and when I give 1 its gives 14 fold change.
>
> dat.m<-read.csv("datam.csv",skip=0)
> genes<-rownames(dat.m)
> groups<-c(2,2,1,1)
> groups<-as.factor(groups)
> design<-model.matrix(~groups)
> design
> fit<-lmFit(dat.m, design)
> fit<-eBayes(fit)
> toptable(fit, coef=2)
>
> Thanks in advance,
> Diya
>
> On Aug 31, 8:26 am, Mathieu Parent <parent.math...@gmail.com> wrote:
>> Salut !
>>
>> Here is the little function I wrote to run limma and save time. I'm a bad
>> scripter, but this saves me a lot of time. Limma is very well documented.
>> Here is a book chapter from Gordon Smith that covers it
>> all:www.statsci.org/smyth/pubs/*limma*-biocbook-reprint.pdf
>>
>> Cheers !
>>
>> (load the function)
>> The call for you would be...
>>
>> data <- matrix_from_limma_not_logged
>> design <- cbind(c(1,1,0,0), c(0,0,1,1))
>> colnames(design) <- c("Gr1", "Gr2")
>> rownames(design) <- c("array1", "array2", "array3", "array4")
>> contrast <- makeContrasts(Gr1VsGr2 = (Gr2-Gr1), levels=design)
>> Pval <- 0.05
>> FDR <- "fdr"
>> LFC <- log2(2)*
>>
>> *output_from_limma <- fun.LIMMA(data, design, contrast, Pval, LFC, FDR)
>> rm(design, contrast, Pval, LFC, FDR)
>>
>> *##### Execute LIMMA to output matrix of significant genes
>> ### ARGUMENTS
>> ## fun.LIMMA(data, design, contrast, Pval, LFC, FDR)
>> # Data matrix - Output from Aroma, NOT LOGGED. rows=genes, cols=arrays
>> # Design matrix for n array in 2 group (ie design <- cbind(c(1,1,0,0),
>> c(0,0,1,1) )
>> # Contrast matrix (ie contrast <- makeContrasts(Gr1VsGr2 = (Gr2-Gr1),
>> levels=design) )
>> # Pval (numeric, between 0 and 1) (ie Pval <- 0.05)
>> # LFC (numeric, logged base 2 value of fold change) (ie LFC<- log2(2))
>> # FDR correction (Caracter, '"none"', '"BH"', '"fdr"', '"BY"' or '"holm" )
>> (ie FDR <- "fdr")
>> *
>> fun.LIMMA <- function(data, design, contrast, Pval, LFC, FDR) {
>> library(limma)
>> # Fit
>> fit.lm <- lmFit(log(data, 2), design)
>> fit.co <- contrasts.fit(fit.lm, contrast)
>> fit.eb <- eBayes(fit.co)
>> # Results
>> results <- decideTests(fit.eb, adjust.method=FDR, p.val=Pval,
>> lfc=LFC)
>> summary_table<-summary(results)
>> # Filtering
>> signif_probes <- which(results != 0)
>> limma_output <- fit.eb[signif_probes,]
>> limma_output$signif_probes <- signif_probes
>> # Output
>> cat("\n", "\n"); cat("*** Input:", "\n") ; cat("\n", "Pval = ",
>> Pval)
>> cat("\n", "LFC = ", LFC)
>> cat("\n", "FDR correction method = ", FDR, "\n", "\n")
>> cat("*** Number of significant probes :", length(signif_probes),
>> "\n", "\n")
>> cat("*** Summary of the results", "\n", "\n");
>> print(summary(results)); cat("\n")
>> cat("*** Object returned: limma_output", "\n", "\n")
>> limma_output #returns the table as an object
>>
>> On Fri, Aug 28, 2009 at 8:13 AM, Diya v <diya_2...@yahoo.co.in> wrote:
>> > Hi
>>
>> > I have 2 control and 2 treatment groups of MoGene-1_0-st.I have the data
>> > normalized and and a data matrix after fit(plm) is performed.
>>
>> > I want to do statistical analysis for differentially expressed ganes.
>>
>> > Can I take the datamatrix generated from aroma.affymetrix and do the
>> > analysis with limma
>>
>> > Is there any online tutorial for this?
>>
>> > Thanks,
>> > Diya
>>
>> > --- On *Fri, 28/8/09, Mathieu Parent <parent.math...@gmail.com>* wrote:
>>
>> > From: Mathieu Parent <parent.math...@gmail.com>
>> > Subject: [aroma.affymetrix] Re: Discussion on gene-1-0-st-array-analysis
>> > To: aroma-affymetrix@googlegroups.com
>> > Date: Friday, 28 August, 2009, 5:58 PM
>>
>> > Hi,
>> > They way it has been proposed to me, is to extract the matrix from the
>> > normalised and summarised data, log it and pass into the LIMMA package for
>> > differential expression analysis.
>>
>> > What is your experimental design ?
>>
>> > Math
>> > McGill University
>>
>> > On Thu, Aug 27, 2009 at 3:46 PM, Diya Vaka
>> > <biotechd...@gmail.com<http://mc/compose?to=biotechd...@gmail.com>
>> > > wrote:
>>
>> >> Hello All,
>>
>> >> I want to know about the up and down regulated genes.So how am i
>> >> supposed to proceed after this step
>>
>> >> Diya
>>
>> > ------------------------------
>> > Love Cricket? Check out live scores, photos, video highlights and more.
>> > Click
>> > here <http://in.rd.yahoo.com/tagline_cricket_2/*http://cricket.yahoo.com>.
> >
>
--~--~---------~--~----~------------~-------~--~----~
When reporting problems on aroma.affymetrix, make sure 1) to run the latest
version of the package, 2) to report the output of sessionInfo() and
traceback(), and 3) to post a complete code example.
You received this message because you are subscribed to the Google Groups
"aroma.affymetrix" group.
To post to this group, send email to aroma-affymetrix@googlegroups.com
To unsubscribe from this group, send email to
aroma-affymetrix-unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/aroma-affymetrix?hl=en
-~----------~----~----~----~------~----~------~--~---
