Hi Lavinia, One reason we changed the default normalization method as quantile is that quantile is more popular method and seems more robust for most cases (other methods should be used if users are care about over-adjustment.) . Also, we are still improving the RSN normalization method and will release the updated version very soon. In order to let users know what method is used, the functions (in developing version) will print out the method in use. Thanks for reporting this! Pan On 6/25/08 6:57 PM, "Lavinia Gordon" <lavinia.gordon@mcri.edu.au> wrote: > Hi Pan > > Many thanks for the prompt reply to my email, that is greatly appreciated. > As you have changed the default normalization, is this because you now feel > that quantile is more appropriate, or is it because quantile is the more > commonly encountered normalization method? > > Do you feel that it would be worth putting a note out about this on the > Bioconductor mailing list? I re-read the vignettes and notice that you do > mention the change there, but there is nothing in the R help about a > (different) default method for lumiN. > > with regards > > Lavinia. > > At 09:36 AM 25/06/2008 -0500, you wrote: >> Hi Lavinia, >> >> We changed the default normalization method as quantile (1.6.0) from previous >> rsn (lumi 1.4.0). This might be the reason of different result. My suggest >> is to keep the processed data (.Rdata) together with the code, so you can >> avoid the difference between different versions of packages. Thanks! >> >> Have a nice day, >> >> >> Pan >> >> >> On 6/25/08 12:56 AM, "Lavinia Gordon" <lavinia.gordon@mcri.edu.au> wrote: >> Dear Dr Pan >> >> I am just confirming that nothing has changed in the most recent version of >> lumi with regard to normalization, background correction or transformation. >> The reason that I ask is that I have just reanalysed a mouse Illumina array >> and the exported datamatrix differs to the one that I previously exported. >> It is quite possible that I have made a mistake somewhere, but as I am using >> exactly the same code I cannot see where this might have occurred. My code >> is: >> >> library(lumi) >> >> # specify the file name >> >> fileName <- 'iltp1444_Sample_Probe_Profile.txt' >> >> # load the data >> >> x.lumi <- lumiR(fileName) >> >> # check it x.lumi # visual check summary(x.lumi, "QC") >> >> # x.lumi = lumiB(x.lumi, method='bgAdjust', probs=0) # [1] "There is no >> control probe information in the LumiBatch object!\n No background adjustment >> will be performed." >> >> x.lumi = lumiB(x.lumi) >> >> # Transfer the Illumina data to stabilize the variance: vst lumi.T <- >> lumiT(x.lumi) # VST transform >> >> lumi.N <- lumiN(lumi.T) # RSN normalization, recommended by lumi >> >> lumi.N.Q <- lumiQ(lumi.N) # visual check # lumi_summary <- summary(lumi.N.Q, >> "QC") >> >> dataMatrix <- exprs(lumi.N) write.csv(DataMatrix, file="dataMatrix.csv") >> >> >> >>> > sessionInfo() R version 2.7.0 (2008-04-22) i386-pc-mingw32 >> >> locale: >> LC_COLLATE=English_Australia.1252;LC_CTYPE=English_Australia.1252;L C_MONETARY >> =English_Australia.1252;LC_NUMERIC=C;LC_TIME=English_Australia.1252 >> >> attached base packages: [1] tools stats graphics grDevices utils >> datasets methods base >> >> other attached packages: [1] lumiMouseAll.db_1.2.0 lumiHumanAll.db_1.2.0 >> AnnotationDbi_1.2.1 RSQLite_0.6-8 DBI_0.2-4 [6] lumi_1.6.0 >> mgcv_1.3-30 affy_1.18.1 preprocessCore_1.2.0 affyio_1.8.0 >> [11] Biobase_2.0.1 > >> >> with many thanks for your time, >> >> Lavinia Gordon. >> >> >> >> ------------------------------------------------------- Bioinformatics >> Department Rm SW916, 9th Floor Murdoch Children's Research Institute Royal >> Children's Hospital Flemington Road, Parkville Victoria 3052 Australia Tel: >> (+613) 8341 6221 ------------------------------------------------------- >> >> >> ------------------------------------------------------ >> Pan Du, PhD >> Research Assistant Professor >> Northwestern University Biomedical Informatics Center >> 750 N. Lake Shore Drive, 11-176 >> Chicago, IL 60611 >> Office (312) 503-2360; Fax: (312) 503-5388 >> dupan@northwestern.edu >> ------------------------------------------------------ >> > > ------------------------------------------------------- > Bioinformatics Department > Rm SW916, 9th Floor > Murdoch Children's Research Institute > Royal Children's Hospital > Flemington Road, Parkville > Victoria 3052 > Australia > Tel: (+613) 8341 6221 > ------------------------------------------------------- [[alternative HTML version deleted]]