I am analyzing paired samples (RNA-seq). I am facing problem with edgeR at y <- estimateGLMTrendedDisp(d, design) step where it is saying error message "Error in t(y) + prior.count.scaled : non-conformable arrays" although it was ok with the step y <- estimateGLMCommonDisp(d, design) Apparently in the assignment step something is happening [design <- model.matrix (~Patient+Tissue)] and when I am trying to see whats there in store at Design, its showing the matrix leaving the one of the sample column absent. Is that exactly what I should expect? Or this is the reason behind the error I am facing. Any suggestion is welcome. Thanks Navonil De Sarkar Senior Research Fellow Human Genetics Unit Indian Statistical Institute 203 B.T. Road Kolkata 700108 Mob:+91 9088153412 Skype. navonil.de.sarkar E.mail ID. navonil.de@gmail.com "I am all in favor of keeping dangerous weapons out of hands of Dumboes Lets start with" Key board [[alternative HTML version deleted]]

Hi, On Wed, Feb 5, 2014 at 7:23 AM, Navonil De Sarkar <navonil.de at="" gmail.com=""> wrote: > I am analyzing paired samples (RNA-seq). I am facing problem with edgeR at > y <- estimateGLMTrendedDisp(d, design) step where it is saying error > message "Error in t(y) + prior.count.scaled : non-conformable arrays" > > although it was ok with the step y <- estimateGLMCommonDisp(d, design) > > Apparently in the assignment step something is happening [design <- > model.matrix (~Patient+Tissue)] and when I am trying to see whats there in > store at Design, its showing the matrix leaving the one of the sample > column absent. Could you be more specific? If your design matrix isn't too big, perhaps you can show it, as well as the levels of Patient and Tissue. My guess is that the sample that looks to be "absent" is likely just absorbed as the (Intercept). -steve -- Steve Lianoglou Computational Biologist Genentech

Dear Navonil de Sarkar, I suspect that you may be using an old version of the software. If so, please upgrade to the current Bioconductor release. If you still have a problem, please give more information about the code sequence leading up to the error message. A read of the posting guide may help: http://www.bioconductor.org/help/mailing-list/posting-guide/ Best wishes Gordon > Date: Wed, 5 Feb 2014 20:53:54 +0530 > From: Navonil De Sarkar <navonil.de at="" gmail.com=""> > To: <bioconductor at="" stat.math.ethz.ch=""> > Subject: [BioC] Error messege in edgeR: Error in t(y) + > prior.count.scaled : non-conformable arrays > > I am analyzing paired samples (RNA-seq). I am facing problem with edgeR at > y <- estimateGLMTrendedDisp(d, design) step where it is saying error > message "Error in t(y) + prior.count.scaled : non-conformable arrays" > > although it was ok with the step y <- estimateGLMCommonDisp(d, design) > > Apparently in the assignment step something is happening [design <- > model.matrix (~Patient+Tissue)] and when I am trying to see whats there in > store at Design, its showing the matrix leaving the one of the sample > column absent. > > Is that exactly what I should expect? Or this is the reason behind the > error I am facing. Any suggestion is welcome. > > Thanks > > > > Navonil De Sarkar > Senior Research Fellow > Human Genetics Unit > Indian Statistical Institute > 203 B.T. Road Kolkata 700108 > Mob:+91 9088153412 > Skype. navonil.de.sarkar > E.mail ID. navonil.de at gmail.com > > "I am all in favor of keeping dangerous weapons out of hands of Dumboes > Lets start with" Key board ______________________________________________________________________ The information in this email is confidential and intend...{{dropped:4}}