Hi, I was wondering how to create my experimental design for my DESeq analysis I have three conditions (ctrl, KO1 , KO2) of three type (I, NP, P) (total - 9 different comaprisons) my phenoData file looks like that SampleName condition libType G1_I_C ctrl I G1_I_KO1 KO1 I G1_I_KO2 KO2 I G1_NP_C ctrl NP G1_NP_KO1 KO1 NP G1_NP_KO2 KO2 NP G1_P_C ctrl P G1_P_KO1 KO1 P G1_P_KO2 KO2 P G2_I_C ctrl I G2_I_KO1 KO1 I G2_I_KO2 KO2 I G2_NP_C ctrl NP G2_NP_KO1 KO1 NP G2_NP_KO2 KO2 NP G2_P_C ctrl P G2_P_KO1 KO1 P G2_P_KO2 KO2 P G3_I_C ctrl I G3_I_KO1 KO1 I G3_I_KO2 KO2 I G3_NP_C ctrl NP G3_NP_KO1 KO1 NP G3_NP_KO2 KO2 NP G3_P_C ctrl P G3_P_KO1 KO1 P G3_P_KO2 KO2 P I would like to compare for each of the three lib_type the differentially regulated genes between KO1/KO2 and the ctrl. Is it possible to do a pair-wise comparison to see the differences in each pair if I create a multifacotrial design matrix like that: fit1 = fitNbinomGLMs( cdsFull, count ~ condition + libType ) fit0 = fitNbinomGLMs( cdsFull, count ~ condition ) or do I need to run each comparison separately like that: res = nbinomTest( cds, group1, group2 ) where as group1 and group2 are each time the ctrl and KO1/KO2 for each libType. Thanks Assa > sessionInfo() R version 3.0.1 (2013-05-16) Platform: x86_64-apple-darwin10.8.0 (64-bit) locale: [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8 attached base packages: [1] parallel stats graphics grDevices utils datasets methods [8] base other attached packages: [1] RColorBrewer_1.0-5 genefilter_1.42.0 DESeq_1.12.1 lattice_0.20-24 [5] locfit_1.5-9.1 Biobase_2.20.1 BiocGenerics_0.6.0 loaded via a namespace (and not attached): [1] annotate_1.38.0 AnnotationDbi_1.22.6 DBI_0.2-7 [4] geneplotter_1.38.0 grid_3.0.1 IRanges_1.18.4 [7] RSQLite_0.11.4 splines_3.0.1 stats4_3.0.1 [10] survival_2.37-4 tools_3.0.1 XML_3.95-0.2 [13] xtable_1.7-1 [[alternative HTML version deleted]]