Bonjour Mathieu. A few comments below.
On 30/07/2009, at 4:35 AM, Mathieu wrote: > > Salut !! > > I am analysing rat exon arrays from Affymetrix with Aroma.Affymetrix > (sessionInfo() bellow). > > What I want, is a differential expression analysis between my two > groups. > > Is extracting the matrix of plmTr and pass a log2() of that to LIMMA > the right thing to do ? I am asking because it seems to be what people > are doing to do such analysis here. Yep. That seems like a reasonably 'standard' thing to do. > But by doing so, aren't we losing > some analysis power by losing the different statistics involved by the > fact that all those probes in a transcript may have different > behaviors ? and all those of an exon ? > > I was proposed to run LIMMA on all the probes of the array, get the t > statistics out of it and then using a Wilcox Ranked test to compare > each units to the distribution of all the t statistics... That's > feasible for me but will involve a lot of head scratching :) You *could* do any number of things. But, you would have to justify all these steps and demonstrate that it does "better" than the standard methods. That is generally difficult to do. > Honestly, I am not a very good statistician (yet!) and a begginner > programmer and I want the safe way to have my differential expression > between my two groups, done using good statistics. Seems like a good starting point would be RMA (or GCRMA) at the gene- level and a limma analysis afterwards. > Second question. I don't get anything out of my analysis if I use a > FDR correction. I thought that would be ok if I used the core cdf > only, but it seems to not be the case. Nothing is significant between > my two groups.. :/ > Is the following the right thing to do ? > ############################################ > method <- "fdr" > pval <- 0.05 > lfc <- 1 # log2(2) > results <- decideTests(fit.eb, adjust.method=method, p.val=pval, > lfc=lfc) > ############################################ This all seems pretty standard. With respect to the fact that nothing is signfificant, that can be due a number of things: data quality, small sample size, true differences are very small. Hope that helps. Cheers, Mark > Thanks a lot in advance ! If needed, I could forward the whole code > but it's basically the user case example of the Human exon array. > > Merci, > Mathieu > McGill University > ********************************************************************* >> sessionInfo() > R version 2.8.1 (2008-12-22) > x86_64-unknown-linux-gnu > > attached base packages: > [1] stats graphics grDevices utils datasets methods > base > > other attached packages: > [1] plotrix_2.5-2 limma_2.16.4 > aroma.affymetrix_1.1.1 > [4] aroma.apd_0.1.6 affxparser_1.14.2 > R.huge_0.1.8 > [7] aroma.core_1.1.2 aroma.light_1.12.2 > matrixStats_0.1.6 > [10] R.rsp_0.3.4 R.filesets_0.5.2 > digest_0.3.1 > [13] R.cache_0.1.7 R.utils_1.1.7 > R.oo_1.4.8 > [16] EBImage_2.6.0 R.methodsS3_1.0.3 > > > ------------------------------ Mark Robinson, PhD (Melb) Epigenetics Laboratory, Garvan Bioinformatics Division, WEHI e: m.robin...@garvan.org.au e: mrobin...@wehi.edu.au p: +61 (0)3 9345 2628 f: +61 (0)3 9347 0852 ------------------------------ --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---