Question: DESeq estimateDispersions
0
gravatar for Lucia Peixoto
6.3 years ago by
Lucia Peixoto330
Lucia Peixoto330 wrote:
Hi, I am trying to run DESeq and I keep getting the following error after trying estimateDispersions Error in parametricDispersionFit(means, disps) : Parametric dispersion fit failed. Try a local fit and/or a pooled estimation. (See '?estimateDispersions') I cannot figure out why this is happening, I have 4 control and 4 case samples the code immediately preceding the error looks like this; conds=rep(0,ncol(C)) cds <- newCountDataSet( C, conds ) cds <- estimateSizeFactors( cds ) cds <- estimateDispersions( cds ) where C is my coverage matrix The matrix does have 355042 rows, is that a problem? thanks for the help PS: This is not "traditional" RNA-seq data, each row is a window in the genome and the values on the matrix are mapped reads in each sample for each window -- Lucia Peixoto PhD Postdoctoral Research Fellow Laboratory of Dr. Ted Abel Department of Biology School of Arts and Sciences University of Pennsylvania "Think boldly, don't be afraid of making mistakes, don't miss small details, keep your eyes open, and be modest in everything except your aims." Albert Szent-Gyorgyi [[alternative HTML version deleted]]
coverage deseq • 2.3k views
ADD COMMENTlink modified 6.3 years ago by Simon Anders3.5k • written 6.3 years ago by Lucia Peixoto330
Answer: DESeq estimateDispersions
0
gravatar for Michael Love
6.3 years ago by
Michael Love24k
United States
Michael Love24k wrote:
hi Lucia, On Wed, Apr 17, 2013 at 5:32 PM, Lucia Peixoto <luciap@iscb.org> wrote: > Hi, > > I am trying to run DESeq and I keep getting the following error after > trying estimateDispersions > > Error in parametricDispersionFit(means, disps) : > Parametric dispersion fit failed. Try a local fit and/or a pooled > estimation. (See '?estimateDispersions') > > > I cannot figure out why this is happening, I have 4 control and 4 case > samples > the code immediately preceding the error looks like this; > conds=rep(0,ncol(C)) > cds <- newCountDataSet( C, conds ) > cds <- estimateSizeFactors( cds ) > cds <- estimateDispersions( cds ) > > You have set all the conditions equal to zero. This should instead be a factor which gives the condition of the samples. See the description of this in ?newCountDataSet. If you continue to get the error, follow the instructions in the message, i.e. read the help for ?estimateDispersions and try using the argument fitType="local" > where C is my coverage matrix > The matrix does have 355042 rows, is that a problem? > > The number of rows is not a problem. Mike [[alternative HTML version deleted]]
ADD COMMENTlink written 6.3 years ago by Michael Love24k
Answer: DESeq estimateDispersions
0
gravatar for Simon Anders
6.3 years ago by
Simon Anders3.5k
Zentrum für Molekularbiologie, Universität Heidelberg
Simon Anders3.5k wrote:
Hi On 17/04/13 17:32, Lucia Peixoto wrote: > Error in parametricDispersionFit(means, disps) : > Parametric dispersion fit failed. Try a local fit and/or a pooled > estimation. (See '?estimateDispersions') [...] > PS: This is not "traditional" RNA-seq data, each row is a window in the > genome and the values on the matrix are mapped reads in each sample for > each window The parametric fit fails sometimes, especially on non-standard data sets. This is why we offer the local fit as a fall-back. Simon
ADD COMMENTlink written 6.3 years ago by Simon Anders3.5k
HI All, thanks very much for the help, doing local fit fixes the issue I a not sure I understand what is the impact of doing this in terms of the accuracy of the estimates can someone explain? will it affect greatly the number of differential calls at a particular cutoff? thanks Lucia On Wed, Apr 17, 2013 at 1:41 PM, Simon Anders <anders@embl.de> wrote: > Hi > > > On 17/04/13 17:32, Lucia Peixoto wrote: > >> Error in parametricDispersionFit(means, disps) : >> Parametric dispersion fit failed. Try a local fit and/or a pooled >> estimation. (See '?estimateDispersions') >> > [...] > > PS: This is not "traditional" RNA-seq data, each row is a window in the >> genome and the values on the matrix are mapped reads in each sample for >> each window >> > > The parametric fit fails sometimes, especially on non-standard data sets. > This is why we offer the local fit as a fall-back. > > Simon > > > ______________________________**_________________ > Bioconductor mailing list > Bioconductor@r-project.org > https://stat.ethz.ch/mailman/**listinfo/bioconductor<https: stat.et="" hz.ch="" mailman="" listinfo="" bioconductor=""> > Search the archives: http://news.gmane.org/gmane.** > science.biology.informatics.**conductor<http: news.gmane.org="" gmane.="" science.biology.informatics.conductor=""> > -- Lucia Peixoto PhD Postdoctoral Research Fellow Laboratory of Dr. Ted Abel Department of Biology School of Arts and Sciences University of Pennsylvania "Think boldly, don't be afraid of making mistakes, don't miss small details, keep your eyes open, and be modest in everything except your aims." Albert Szent-Gyorgyi [[alternative HTML version deleted]]
ADD REPLYlink written 6.3 years ago by Lucia Peixoto330
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 129 users visited in the last hour