Dear Dr. Michael, I???m using DESeq2 to compare data from different treatments to find possible bioindicators. I applied the DESeq function by two ways: A) treatmentdeseq <- phyloseq_to_deseq2(biom_otu_tax, ~treatment) B) treatmentdeseq2<- phyloseq_to_deseq2(biom_otu_tax, ~treatment + condition x) The intercept in both is the same, but the comparisons among treatments are (a bit) different when I call A and B. In B, does the condition exert some influence on analysis comparing treatments? Could you explain why? Thanks, Leonardo M. Pitombo -- output of sessionInfo(): > sessionInfo("DESeq2") R version 3.0.1 (2013-05-16) Platform: x86_64-w64-mingw32/x64 (64-bit) locale: [1] LC_COLLATE=Dutch_Netherlands.1252 LC_CTYPE=Dutch_Netherlands.1252 LC_MONETARY=Dutch_Netherlands.1252 [4] LC_NUMERIC=C LC_TIME=Dutch_Netherlands.1252 attached base packages: character(0) other attached packages: [1] DESeq2_1.2.10 loaded via a namespace (and not attached): [1] ade4_1.6-2 annotate_1.40.0 AnnotationDbi_1.24.0 ape_3.0-11 [5] base_3.0.1 Biobase_2.22.0 BiocGenerics_0.8.0 biom_0.3.11 [9] Biostrings_2.30.1 cluster_1.14.4 codetools_0.2-8 colorspace_1.2-4 [13] datasets_3.0.1 DBI_0.2-7 dichromat_2.0-0 digest_0.6.4 [17] foreach_1.4.1 genefilter_1.44.0 GenomicRanges_1.14.4 ggplot2_0.9.3.1 [21] glmmADMB_0.7.7 graphics_3.0.1 grDevices_3.0.1 grid_3.0.1 [25] gtable_0.1.2 igraph_0.7.0 IRanges_1.20.6 iterators_1.0.6 [29] labeling_0.2 lattice_0.20-24 locfit_1.5-9.1 MASS_7.3-29 [33] Matrix_1.1-2 methods_3.0.1 multtest_2.18.0 MuMIn_1.9.13 [37] munsell_0.4.2 nlme_3.1-113 parallel_3.0.1 permute_0.8-3 [41] phyloseq_1.7.12 plyr_1.8 proto_0.3-10 R2admb_0.7.10 [45] RColorBrewer_1.0-5 Rcpp_0.11.0 RcppArmadillo_0.4.000.2 reshape2_1.2.2 [49] RJSONIO_1.0-3 RSQLite_0.11.4 scales_0.2.3 splines_3.0.1 [53] stats_3.0.1 stats4_3.0.1 stringr_0.6.2 survival_2.37-7 [57] tools_3.0.1 utils_3.0.1 vegan_2.0-10 XML_3.98-1.1 [61] xtable_1.7-1 XVector_0.2.0 -- Sent via the guest posting facility at bioconductor.org.

hi Leonardo, (side note: we recommend to put the variable of interest always at the end of the design formula, so the defaults for results() and plotMA() give you the contrast of interest) On Tue, Mar 4, 2014 at 11:44 AM, Leonardo Pitombo [guest] < guest@bioconductor.org> wrote: > > Dear Dr. Michael, > > I’m using DESeq2 to compare data from different treatments to find > possible bioindicators. I applied the DESeq function by two ways: > > A) treatmentdeseq <- phyloseq_to_deseq2(biom_otu_tax, ~treatment) > > B) treatmentdeseq2<- phyloseq_to_deseq2(biom_otu_tax, ~treatment + > condition x) > > The intercept in both is the same, but the comparisons among treatments > are (a bit) different when I call A and B. In B, does the condition exert > some influence on analysis comparing treatments? > Could you explain why? > ​The treatment effect is different in (B) because we say the condition effect has been "accounted for".​ Try to find a reference on linear modeling, as the change in estimated coefficients after "accounting for" other variables is a general property of linear models. For example, compare the estimate for x in these two linear models: > x <- rep(0:1, each=20) > z <- c(rep(0:1,c(15,5)),rep(0:1,c(5,15))) say x is our variable of interest, and we have an unbalanced batch variable z: > z [1] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 [39] 1 1 > y <- rnorm(40, x + 2*z) > coef(summary(lm(y ~ x))) Estimate Std. Error t value Pr(>|t|) (Intercept) 0.4343922 0.2601761 1.669608 1.032160e-01 x 2.0148582 0.3679446 5.475982 2.978237e-06 > coef(summary(lm(y ~ z + x))) Estimate Std. Error t value Pr(>|t|) (Intercept) -0.06667297 0.1833274 -0.3636824 7.181651e-01 z 2.00426063 0.2771650 7.2312893 1.397230e-08 x 1.01272784 0.2771650 3.6538801 7.959908e-04 > ​ ​ > > Thanks, > Leonardo M. Pitombo > > > -- output of sessionInfo(): > > > sessionInfo("DESeq2") > R version 3.0.1 (2013-05-16) > Platform: x86_64-w64-mingw32/x64 (64-bit) > > locale: > [1] LC_COLLATE=Dutch_Netherlands.1252 LC_CTYPE=Dutch_Netherlands.1252 > LC_MONETARY=Dutch_Netherlands.1252 > [4] LC_NUMERIC=C LC_TIME=Dutch_Netherlands.1252 > > attached base packages: > character(0) > > other attached packages: > [1] DESeq2_1.2.10 > > loaded via a namespace (and not attached): > [1] ade4_1.6-2 annotate_1.40.0 AnnotationDbi_1.24.0 > ape_3.0-11 > [5] base_3.0.1 Biobase_2.22.0 BiocGenerics_0.8.0 > biom_0.3.11 > [9] Biostrings_2.30.1 cluster_1.14.4 codetools_0.2-8 > colorspace_1.2-4 > [13] datasets_3.0.1 DBI_0.2-7 dichromat_2.0-0 > digest_0.6.4 > [17] foreach_1.4.1 genefilter_1.44.0 GenomicRanges_1.14.4 > ggplot2_0.9.3.1 > [21] glmmADMB_0.7.7 graphics_3.0.1 grDevices_3.0.1 > grid_3.0.1 > [25] gtable_0.1.2 igraph_0.7.0 IRanges_1.20.6 > iterators_1.0.6 > [29] labeling_0.2 lattice_0.20-24 locfit_1.5-9.1 > MASS_7.3-29 > [33] Matrix_1.1-2 methods_3.0.1 multtest_2.18.0 > MuMIn_1.9.13 > [37] munsell_0.4.2 nlme_3.1-113 parallel_3.0.1 > permute_0.8-3 > [41] phyloseq_1.7.12 plyr_1.8 proto_0.3-10 > R2admb_0.7.10 > [45] RColorBrewer_1.0-5 Rcpp_0.11.0 > RcppArmadillo_0.4.000.2 reshape2_1.2.2 > [49] RJSONIO_1.0-3 RSQLite_0.11.4 scales_0.2.3 > splines_3.0.1 > [53] stats_3.0.1 stats4_3.0.1 stringr_0.6.2 > survival_2.37-7 > [57] tools_3.0.1 utils_3.0.1 vegan_2.0-10 > XML_3.98-1.1 > [61] xtable_1.7-1 XVector_0.2.0 > > -- > Sent via the guest posting facility at bioconductor.org [[alternative HTML version deleted]]