Hi Xinjun. Comments below.
On 28/04/2009, at 12:25 PM, Xinjun Zhang wrote: > Hi Mark: > > Thanks very much for your clarification! Now I have approached to > limma analysis of FIRMA score to get differentially spliced genes > ( and also splicing pattern of each ). But I still have some > difficulty to understand the code ( in red ) below in Limma analysis: > > #fs is the 'standard' FirmaSet-object > fsDF <- extractDataFrame(fs, addNames=TRUE) > fsDF[,-c(1:5)] <- log2(fsDF[,-c(1:5)]) # I know why log2 > is here but confused by fsDF[,c(1:5)] .... what does this expression > mean? Note that it is -c(1:5), meaning operate on (here, take logs) all of the columns except 1:5 ... that is, because extractDataFrame gives some extra columns at the beginning that are NOT data, we only want to log that columns that have actual data. > design <- cbind(Grp1=1,Grp2=c(rep(0,n_1),rep(1,n_2))) > fit<-lmFit(fsDF[,-c(1:5)],design) > fit<-eBayes(fit) > fit$genes<-fsDF[,1] # Can I also get seperate splicing > patterns for the two differentially spliced genes from two group > (control and treatment )? I'm not sure what you are asking here. The probesets where the "Grp2" coefficient is significantly different from 0 may highlight differentially spliced exons. Does that help? Mark > Thanks in advance! > > Xinjun > > On Mon, Apr 27, 2009 at 6:19 AM, Mark Robinson > <mrobin...@wehi.edu.au> wrote: > > > Hi Xinjun. > > Quick comments below. > > > > Hi Mark: > > > > Thanks very much for your help and I have have got a quick start > on a > > small > > dataset that each group (control and treatment ) contains 4 > arrays. I have > > set up a file structure like this: > > ================================================= > > rawDate/ > > controlGroup/ > > HuEx-1_0-st-v1/ > > GSMXXXXXX.CEL > > GSMXXXXXX.CEL > > ............................ > > > > treatmentGroup/ > > HuEx-1_0-st-v1/ > > GSMXXXXXX.CEL > > GSMXXXXXX.CEL > > ............................ > > ==================================================* > > > This setup will need to be changed. You will want to put ALL samples > together to do the PLM fitting, normalization, FIRMA scoring, etc. > > Something like: > > rawData/ > thisExperiment/ > HuEx-1_0-st-v1/ > sample1.CEL > sample2.CEL > ... > sampleN.CEL > > > > This is my code ( my questions are in red):* > > > > library(aroma.affymetrix) > > > > #Getting annotation data files > > chipType <- "HuEx-1_0-st-v1" > > cdf <- AffymetrixCdfFile$byChipType(chipType) > > print(cdf) > > > > #Defining CEL set > > cs <- AffymetrixCelSet$byName("controlGroup", cdf=cdf) > > print(cs) > > > > #Background Adjustment and Normalization > > bc <- RmaBackgroundCorrection(cs) > > csBC <- process(bc,verbose=verbose) > > > > #quantile normalization > > qn <- QuantileNormalization(csBC, typesToUpdate="pm") ### I set the > > second > > parameter as "pm" as the chip type is Affymetrix exon array, is that > > right? > > print(qn) > > csN <- process(qn, verbose=verbose) > > This is fine. > > > > > #Summarization > > getCdf(csN) > > ## * Fit exon-by-exon*, change the value of mergeGroups to FALSE > in the > > ExonRmaPlm() call above. > > *plmEx *<- ExonRmaPlm(csN, mergeGroups=*FALSE*) > > print(*plmEx*) > > #To fit the PLM to all of the data, do: > > fit(*plmEx*, verbose=verbose) > > * > > And here is my problem:* > > firma <- FirmaModel(plmTr) # I have noticd that FIRMA analysis > ONLY works > > from the PLM based on transcripts. So when the parameter is plmTr, I > > wonder > > how can it detect the splicing events of genes ? Should not the > parameter > > be > > plmEx? > > fit(firma, verbose=verbose) > > fs <- getFirmaScores(firma) > > Like it says on the group web page for Exon arrays: "The FIRMA > analysis > ONLY works from the PLM based on transcripts". This is NOT an error. > That's the way it works. > > The manuscript gives more details for why this is the case: > http://bioinformatics.oxfordjournals.org/cgi/content/abstract/24/15/1707 > > Hope that helps. > > Cheers, > Mark > > > > > > > > > > > > > On Fri, Apr 24, 2009 at 5:24 PM, Mark Robinson > > <mrobin...@wehi.edu.au>wrote: > > > >> > >> Hi Xinjun. > >> > >> Here is a quick sketch of what I might do. > >> > >> 1. Run everything to get FIRMA scores. See group page for running > >> details and the Purdom Bioinformatics 2008 paper for methodological > >> details. > >> > >> 2a. If Nn or Nc > 1, use 'limma' to look for a difference in FIRMA > >> scores between your two groups. See threads: > >> > >> http://groups.google.com/group/aroma-affymetrix/browse_thread/thread/36d8c59d742fc503/ > >> > >> http://groups.google.com/group/aroma-affymetrix/browse_thread/thread/7d2645bd76cc2023/ > >> > >> 2b. If you have say patient samples (and a good number of them), > you > >> might expect only a subset of your C or N patients to have a > splicing > >> aberration. In this case, maybe you just want to look for large- > in- > >> magnitude FIRMA scores. > >> > >> ... maybe you also want to look at interesting probesets via > >> GenomeGraphs: > >> > >> http://groups.google.com/group/aroma-affymetrix/web/using-the-genomegraphs-package-with-firma > >> > >> Cheers, > >> Mark > >> > >> > >> On 24/04/2009, at 1:01 AM, liszt wrote: > >> > >> > > >> > Hi: > >> > > >> > Now I have N CEL files from both normal and cancer tissue. The > two > >> > groups contains Nn and Nc CEL files separately (Nn = Nc). I > want to > >> > investigate the difference in gene's splicing pattern between > normal > >> > tissue and cancerous tissue. Would you give me some tips? ( I > have > >> > read the document and got no answers) Thanks! > >> > > >> > Xinjun > >> > > > >> > >> ------------------------------ > >> Mark Robinson > >> 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 > >> ------------------------------ > >> > >> > >> > >> > >> > >> > > >> > > > > > > > > > > > > > > > > ------------------------------ Mark Robinson 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 -~----------~----~----~----~------~----~------~--~---