Dear List, The experiment I have has the following description table: Treatment SampelingPoint Subjects SampleNames placebo_20 20 1 placebo_20_01 placebo_22 22 1 placebo_22_01 celltype2_20 20 2 celltype2_20_02 celltype2_22 22 2 celltype2_22_02 celltype2_20 20 3 celltype2_20_03 celltype2_22 22 3 celltype2_22_03 celltype2_20 20 4 celltype2_20_04 celltype2_22 22 4 celltype2_22_04 placebo_20 20 5 placebo_20_05 placebo_22 22 5 placebo_22_05 celltype2_20 20 7 celltype2_20_07 celltype2_22 22 7 celltype2_22_07 celltype2_20 20 8 celltype2_20_08 celltype2_22 22 8 celltype2_22_08 placebo_20 20 9 placebo_20_09 placebo_22 22 9 placebo_22_09 placebo_20 20 11 placebo_20_11 placebo_22 22 11 placebo_22_11 placebo_20 20 12 placebo_20_12 placebo_22 22 12 placebo_22_12 placebo_20 20 13 placebo_20_13 placebo_22 22 13 placebo_22_13 placebo_20 20 14 placebo_20_14 placebo_22 22 14 placebo_22_14 placebo_20 20 15 placebo_20_15 placebo_22 22 15 placebo_22_15 placebo_20 20 16 placebo_20_16 placebo_22 22 16 placebo_22_16 celltype2_20 20 17 celltype2_20_17 celltype2_22 22 17 celltype2_22_17 placebo_20 20 18 placebo_20_18 placebo_22 22 18 placebo_22_18 celltype2_20 20 19 celltype2_20_19 celltype2_22 22 19 celltype2_22_19 celltype2_20 20 20 celltype2_20_20 celltype2_22 22 20 celltype2_22_20 celltype2_20 20 21 celltype2_20_21 celltype2_22 22 21 celltype2_22_21 celltype2_20 20 24 celltype2_20_24 celltype2_22 22 24 celltype2_22_24 celltype2_20 20 25 celltype2_20_25 celltype2_22 22 25 celltype2_22_25 placebo_20 20 26 placebo_20_26 placebo_22 22 26 placebo_22_26 placebo_20 20 27 placebo_20_27 placebo_22 22 27 placebo_22_27 celltype2_20 20 28 celltype2_20_28 celltype2_22 22 28 celltype2_22_28 celltype2_20 20 29 celltype2_20_29 celltype2_22 22 29 celltype2_22_29 placebo_20 20 30 placebo_20_30 placebo_22 22 30 placebo_22_30 placebo_20 20 31 placebo_20_31 placebo_22 22 31 placebo_22_31 placebo_20 20 32 placebo_20_32 placebo_22 22 32 placebo_22_32 celltype2_20 20 33 celltype2_20_33 celltype2_22 22 33 celltype2_22_33 celltype2_20 20 34 celltype2_20_34 celltype2_22 22 34 celltype2_22_34 I want to see the effect of the samplingPoint within a treatment, because we give the subjects a substance in between day 20 and day 22. In other words, we are interested in the effect of the substance (comparing day 20 versus day 22). At both samplingPoints (day 20 and day 22) 1 sample was drawn. I want to give the analyses extra power by taking into account that the comparison between day 20 and 22 can be made within the same individual (1 individual is measured twice). So my comparisons are: - "placebo_20 – placebo_22" - "celltype2_20 – celltype2_22" I am using 8.3 Paired Samples of the limma user guide as my guide for the limma design setup. library("limma") # Give levels. Will be the names (the unique names of the bio.replicates) of the columns lev <- c( sort(as.character(unique(description$Treatment)))) > lev [1] "substance_20" "substance_22" "placebo_20" "placebo_22" > Subjects <- factor(description$Subjects) > Treatment <- factor(description$Treatment) > design <- model.matrix(~0+Treatment+Subjects) # Name comparison > contrast.fit <- "placebo_20 – placebo_22" design <- model.matrix(~0+Treatment+Subjects) # design N =34 subjects but cutt it off for mailing it here. > design TreatmentVSL3_20 TreatmentVSL3_22 Treatmentplacebo_20 Treatmentplacebo_22 Subjects2 Subjects3 1 0 0 0 1 0 1 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 1 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 If I do a fit on my data using the design matrix above I get the following warning /error > fit <- lmFit(data, design) Coefficients not estimable: Subjects34 Warning message: Partial NA coefficients for 47323 probe(s) Can anyone tell me what's goes wrong my design? Or suggest how to do the limma analyses with my specific experiment. > sessionInfo() R version 2.14.2 (2012-02-29) Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit) locale: [1] C/en_US.UTF-8/C/C/C/C attached base packages: [1] stats graphics grDevices utils datasets methods base other attached packages: [1] qvalue_1.28.0 ggplot2_0.9.1 limma_3.10.3 [4] lumiHumanAll.db_1.16.0 org.Hs.eg.db_2.6.4 lumiHumanIDMapping_1.10.0 [7] RSQLite_0.11.1 DBI_0.2-5 AnnotationDbi_1.16.19 [10] lumi_2.6.0 nleqslv_1.9.3 methylumi_2.0.13 [13] Biobase_2.14.0 loaded via a namespace (and not attached): [1] BiocInstaller_1.2.1 IRanges_1.12.6 KernSmooth_2.23-7 MASS_7.3-18 [5] Matrix_1.0-5 RColorBrewer_1.0-5 affy_1.32.1 affyio_1.22.0 [9] annotate_1.32.3 colorspace_1.1-1 dichromat_1.2-4 digest_0.5.2 [13] grid_2.14.2 hdrcde_2.15 labeling_0.1 lattice_0.20-6 [17] memoise_0.1 mgcv_1.7-16 munsell_0.3 nlme_3.1-103 [21] plyr_1.7.1 preprocessCore_1.16.0 proto_0.3-9.2 reshape2_1.2.1 [25] scales_0.2.1 stringr_0.6 tcltk_2.14.2 tools_2.14.2 [29] xtable_1.7-0 zlibbioc_1.0.1 Thanks a lot, Varshna Goelela [[alternative HTML version deleted]]