Hello List, I am just forwarding Lorenzo's response to the list as he forgot to copy the list. May be, this is of use in the future. I think, your design is a unbalanced randomized incomplete block design, so it doesn't matter if you have one treatment or multiple for a single block. Also, the duplicateCorrelation is estimated from the data and the estimate (0.1313324) being high or low doesn't come into the picture. The data determines this for you, even if it seems that the data from the same mother should be "more correlated". This may happen because of multiple reasons: - May be the treatments are not dependent on the mother - If the treatments are dependent, may be the relation between them is not normal which leads to bad estimate of correlation. and .. several similar reasons .. Bottom line: If you trust the model, the data tells you that 0.1313324 is the estimate of correlation between observations in a block. If you are not happy with the estimate of correlation, probably try fitting a different model. Thanks, S. On Wed, Oct 6, 2010 at 7:30 PM, Lorenzo Bomba wrote: > > > Hi Sunny, > > Thank you for the suggestions, I tried to use duplicate correlation as > follow: > > TS <-target$Treatment > design <- model.matrix(~-1+TS) > colnames(design) <- levels(TS) > block <- c(1,2,3,4,2,3,3,5,5,3,6,6) > dupcor <- duplicateCorrelation(RDE,design,block=block) > > > and the value of dupcor$consensus.correlation is 0.1313324.... really low > for a average correlation. > I could go on with the analysis? as you can see in the block I have some > array correlated 'cause the same mother and other that are independent > without duplicate.... > some other suggestions please?? > > Thank a lot for your effort! > > Cheers > > Lorenzo > > On 6 Oct 2010, at 20:50, Sunny Srivastava wrote: > > Hello Lorenzo, > I am not sure if I understand your question correctly, but here is my > $0.02. > > If you want to account for effect of mother, it would be better include it > in your regression model as a factor, but I guess your aim is to detect the > difference between DR and NODR (right?). If I understand you correctly (ie > you are interested in NODR vs DR comparison), you probably want to model > mothers as blocks. The duplicateCorrelation option in limma might come in > handy in this case. > > Thanks, > S. > > On Wed, Oct 6, 2010 at 12:56 PM, Lorenzo Bomba wrote: > >> Dear all, >> >> I have a conceptual problem with the "makeContrast" function in Limma: >> doing a factorial design and the follow are the targetfile and the >> design matrix >> >> targetfile >> >> FileName Scrofa Treatment >> S10M S127 DR >> S11 S52 DR >> S12 S304 DR >> S13 S9 NODR >> S14 S52 DR >> S15 S304 DR >> S17rt S304 DR >> S18 S66 DR >> S19 S66 NODR >> S20 S304 NODR >> S21 S93 NODR >> S22 S93 NODR >> >> >> > design >> S127.DR S304.DR S304.NODR S52.DR S66.DR S66.NODR S9.NODR S93.NODR >> 1 1 0 0 0 0 0 0 0 >> 2 0 0 0 1 0 0 0 0 >> 3 0 1 0 0 0 0 0 0 >> 4 0 0 0 0 0 0 1 0 >> 5 0 0 0 1 0 0 0 0 >> 6 0 1 0 0 0 0 0 0 >> 7 0 1 0 0 0 0 0 0 >> 8 0 0 0 0 1 0 0 0 >> 9 0 0 0 0 0 1 0 0 >> 10 0 0 1 0 0 0 0 0 >> 11 0 0 0 0 0 0 0 1 >> 12 0 0 0 0 0 0 0 1 >> attr(,"assign") >> [1] 1 1 1 1 1 1 1 1 >> attr(,"contrasts") >> attr(,"contrasts")$TS >> [1] "contr.treatment" >> >> I want see the difference between the NODR and the DR group taking in >> account that they belong to different mothers ( S127 S304 .....etc ). How >> can I set the makeContrast.... >> I tried in this two ways.... but I don't know the difference between the >> first and the second >> >> 1)cont.matrix <- makeContrasts( >> DRvsNODR=(S127.DR+S304.DR+S52.DR)-(S304.NODR+S66.NODR+S9.NODR+S93.NODR), >> levels=design) >> >> 2) cont.matrix <- makeContrasts( >> DRvsNODR=(S127.DR-S304.DR-S52.DR)-(S304.NODR-S66.NODR-S9.NODR-S93.NODR), >> levels=design) >> >> Could you help me please??? >> >> Thanks in advance! >> >> Lorenzo >> >> >> >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor@stat.math.ethz.ch >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: >> http://news.gmane.org/gmane.science.biology.informatics.conductor >> > > > [[alternative HTML version deleted]]