I am working with edgeR to look at differential expression in a 2x2 design of two populations with two treatments (a total of 4 groups with 6 replicates each). I have very interesting results. However, a colleague asked me the other day, "What does edgeR do with zero values for read count data?" I have an intermediate level of statistical understanding at best, and in some instances, a novice level. I am assuming that the negative binomial model in some way accounts for tags with a large number of 0 cells. Is there something more going on the generalized linear model or in fitting the negative binomial distribution that accounts for a large percentage of zeros from some of the tags? Thank you for your guidance! [[alternative HTML version deleted]]

Hi Laura, See ?glmFit, particularly the prior.count artument: prior.count: average prior count to be added to observation to shrink the estimated log-fold-changes towards zero. and this part from the Details section: Positive ?prior.count? cause the returned coefficients to be shrunk in such a way that fold-changes between the treatment conditions are decreased. In particular, infinite fold-changes are avoided. Larger values cause more shrinkage. The returned coefficients are affected but not the likelihood ratio tests or p-values. Best, Jim On Friday, February 07, 2014 1:48:26 PM, Laura Eierman wrote: > I am working with edgeR to look at differential expression in a 2x2 design > of two populations with two treatments (a total of 4 groups with 6 > replicates each). I have very interesting results. However, a colleague > asked me the other day, "What does edgeR do with zero values for read count > data?" > > I have an intermediate level of statistical understanding at best, and in > some instances, a novice level. I am assuming that the negative binomial > model in some way accounts for tags with a large number of 0 cells. Is > there something more going on the generalized linear model or in fitting > the negative binomial distribution that accounts for a large percentage of > zeros from some of the tags? > > Thank you for your guidance! > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099