On Tue, Sep 1, 2009 at 7:53 AM, Henrik Bengtsson<henrik.bengts...@gmail.com> wrote: > On Mon, Aug 31, 2009 at 11:42 PM, ssv<ssv....@gmail.com> wrote: >> >> Hi Henrik >> >> Here are the links: >> >> 1) >> http://groups.google.com/group/aroma-affymetrix/web/total-copy-number-analysis-6-0 >> >> Under "Identification of copy-number regions" >> >> cbs <- CbsModel(ces1, ces2) > > From the vignette and section your referring to: "In order to do > paired analysis using the Circular Binary Segmentation (CBS) method, > do: > > cbs <- CbsModel(ces1, ces2) > > where 'ces1' is a CEL set of test samples and 'ces2' is a CEL set of > the same number of control samples. Use ces1 <- extract(cesN, > arrays1) and same for 'ces2' to extract the two sets from the 'cesN' > CEL set above." > > Whenever you specify two *sets*, you tell the segmentation model that > you want to do a paired/matched analysis. If so, the model calculate > ratios between the first array in ces1 and the first in ces2, between > the seconf in ces1 and the second in ces2, and so on. Therefore, if > 'ces1' and 'ces2' are *sets*, they must contain the same number of > arrays (otherwise an error will be thrown). > >> >> >> 2) In windows, under C:\Program Files\R\R-2.9.1\library >> \aroma.affymetrix\testScripts\system\chipTypes\GenomeWideSNP_6, R >> script file by name "test20080729,6.0,CN,refSet.R" > > FYI, those are not really user example scripts, but the > redundancy/system tests that we run at each release to make sure > everything works as expected and that new updates doesn't break old > functions. Of course, you may still look at them, because most > imitate real scenarios. > >> >> >> # - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - >> - >> # Segmentation with specific reference set >> # - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - >> - >> # Use the robust average of the first three arrays as a reference >> cesR <- extract(cesN, 1:3); >> ceR <- getAverageFile(cesR); >> print(ceR); >> >> sm <- CbsModel(cesN, ceR); > > As print(cesN), print(cesR), and print(ceR) confirms, 'cesN' is a data > *set* with 6 arrays (the "s" in "ces" indicates "set"), 'cesR' is a > subset of 'cesN' with 3 arrays, and 'ceR' is a single array/file (no > "s"). 'ceR' is a single array in the sense that it is a robust > average of the 3 arrays. > > Next, you set up a segmentation model where you want to segment the 6 > arrays in the *set* using one *array* as a reference. This is > interpreted as you want to use the same reference for each of the 6 > arrays. > >> >> >> ===================== >> >> So in my context, i translated that as given below: >> >> Analysis is done on two paired samples (normal1: tumor1). >> >> Script 1 (Link 1) >> ========================= >> # Use the robust average of the first array as a reference >> cesR <- extract(cesN, 1); >> cesL <- extract(cesN, 2); >> sm<-CbsModel(cesR,cesL) > > Note, extract() extracts a *set* of arrays, so you are asking for a > paired analysis. The CN ratios are hence calculated between the first > array in 'cesR' and the first array in 'cesL' (that's all arrays there > is). (FYI, for extracting single arrays there is getFile()). > > >> ========================= >> >> Script 2 (from the link 2: modified example script) >> >> # Use the robust average of the first array as a reference >> cesR <- extract(cesN, 1); >> cesL <- extract(cesN, 2); >> ceR <- getAverageFile(cesR); >> sm <- CbsModel(cesN, ceR); > > 'cesN' is a set (of six arrays from the example) and 'ceR' is an > (robust average) *file*, so no paired analysis but a common reference. > >> ========================= >> >> First question would be: did I cross stitch two different methods >> (analyses) and came up with a hybrid script :(? > > Yes. It is easier to help you if you instead say what you want to do. > ...and using print(cesR) etc to see what is going on.
...from a private followup email message(*): > This almost solved my problem in analysis except for that last part. > Scenario is like this: I have one control sample for multiple tumors > and I wanted them to be compared pairwise. For eg. Normal sample N1, > Tumor samples T1, T2, T3, T4 and I need them to be compared like > this : N1, T1; N1,T2; N1,T3. So, assume that you processed everything together in a set 'cesN' of, say, 12 arrays, and that the tumors have index 3,4, 9 and 11, and the normal has index 2. The you want to do: # The set containing the 4 tumor samples cesTumor <- extract(cesN, c(3,4,9,11)); # The single normal sample (as a reference for all) ceNormal <- getFile(cesN, 2); cbs <- CbsModel(cesTumor, ceNormal); This will process the 4 arrays in 'cesTumor'. The remaining in 'cesN' will not be processed. Hope this helps (*) Please don't send private messages - 'FAQ. 2008-03-26: Why do you not want to answer questions sent to your private email address?': http://groups.google.com/group/aroma-affymetrix/web/frequently-asked-questions Henrik > > Cheers, > > Henrik > >> >> If not, second one would be , what is the difference ? >> >> Regards >> >> suresh >> >> >> >> >> > --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---