Hello everyone, I'm analyzing some RNASeq in R 3.0.0 using DESeq 1.12.0. I need to do a variance stabilizing transformation of my count data so that I can do a Principal Components Analysis. I plotted the distributions first, using plotDispEsts() and I had estimated the dispersions using estimateDispersions() with the default method and those both seemed to work fine. However, when I do estimateDispersions(method="blind") on the same input, I get the following error: "Error in parametricDispersionFit(means, disps) : Parametric dispersion fit failed. Try a local fit and/or a pooled estimation. (See '?estimateDispersions')" I tried using method="pooled", as the message recommends, and that did work, but I'm not completely sure if it is ok to use the pooled results for my PCA or not. I'm looking for advice as to whether I should proceed with the pooled, or if there's something else I need to do to get the blind to work? If I understand the documentation correctly, both blind and pooled, result in a single dispersion value for all samples, so I think I should be ok with pooled, but I wanted to double check that this is the best course of action. Thanks so much. Alex Koeppel P.S. sessionInfo() R version 3.0.0 (2013-04-03) 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] grid parallel stats graphics grDevices utils datasets methods base other attached packages: [1] gplots_2.11.0.1 MASS_7.3-26 KernSmooth_2.23-10 caTools_1.14 [5] gdata_2.12.0.2 gtools_2.7.1 vsn_3.28.0 DESeq_1.12.0 [9] lattice_0.20-15 locfit_1.5-9.1 Biobase_2.20.0 BiocGenerics_0.6.0 [13] BiocInstaller_1.10.1 loaded via a namespace (and not attached): [1] affy_1.38.1 affyio_1.28.0 annotate_1.38.0 AnnotationDbi_1.22.5 [5] bitops_1.0-5 DBI_0.2-7 genefilter_1.42.0 geneplotter_1.38.0 [9] IRanges_1.18.1 limma_3.16.4 preprocessCore_1.22.0 RColorBrewer_1.0-5 [13] RSQLite_0.11.3 splines_3.0.0 stats4_3.0.0 survival_2.37-4 [17] tools_3.0.0 XML_3.95-0.2 xtable_1.7-1 zlibbioc_1.6.0 [[alternative HTML version deleted]]

Hi Alex On 15/07/13 19:33, Alex Koeppel wrote: > However, when I do estimateDispersions(method="blind") on the same input, I > get the following error: > > "Error in parametricDispersionFit(means, disps) : > Parametric dispersion fit failed. Try a local fit and/or a pooled > estimation. (See '?estimateDispersions')" > > I tried using method="pooled", as the message recommends, and that did > work, but I'm not completely sure if it is ok to use the pooled results for > my PCA or not. No, pooled estimation is not appropriate for PCA. You should rather try a local fit (fitType="local"). Simon