Search
Question: Error in estimateDispersionsFit [DESeq2]
0
gravatar for ygchai2015
3.7 years ago by
ygchai20150
Korea, Republic Of
ygchai20150 wrote:

While running DESeq2 on a Human dataset (commands given below), I have noticed the following error. How I can handle my data with this case? Is that the same case of Error in estimateDispersions of DEXSeq? Could you tell me how to solve the problem, please?

> library("DESeq2")
> setwd("/folder/to/path/")
> sampleFiles =grep("count",list.files("/folder/to/path/"),value=TRUE)
> sampleFiles
>  sampleCondition<-c("Cont","Cont","Test-1","Test-1","Test-2","Test-2","Test-3","Test-3")
> sampleTable<-data.frame(sampleName=sampleFiles, fileName=sampleFiles, condition=sampleCondition)
> ddsHTSeq<-DESeqDataSetFromHTSeqCount(sampleTable=sampleTable, directory="/folder/to/path/", design=~condition)
> colData(ddsHTSeq)$condition<-factor(colData(ddsHTSeq)$condition)
> dds<-DESeq(ddsHTSeq)
# estimating size factors
# estimating dispersions
# gene-wise dispersion estimates
# mean-dispersion relationship
# Error in estimateDispersionsFit(object, fitType = fitType, quiet = quiet) : 
#  all gene-wise dispersion estimates are within 2 orders of magnitude
#  from the minimum value, and so the standard curve fitting techniques will not work.
#  One can instead use the gene-wise estimates as final estimates:
#  dds <- estimateDispersionsGeneEst(dds)
#  dispersions(dds) <- mcols(dds)$dispGeneEst

ADD COMMENTlink modified 3.7 years ago • written 3.7 years ago by ygchai20150
0
gravatar for Simon Anders
3.7 years ago by
Simon Anders3.5k
Zentrum für Molekularbiologie, Universität Heidelberg
Simon Anders3.5k wrote:

Yes, this is the same issue.

Are your samples really biological and not technical replicates?

ADD COMMENTlink written 3.7 years ago by Simon Anders3.5k
0
gravatar for ygchai2015
3.7 years ago by
ygchai20150
Korea, Republic Of
ygchai20150 wrote:

Yes. My samples are really biological replicates.   

ADD COMMENTlink written 3.7 years ago by ygchai20150

This means that the replicates are very similar: the coefficient of variation is less than 0.1% for all genes. This is really atypical for biological replicates of RNA-seq experiments. You might look at plotPCA to see how close the replicates are. Also you can plot the counts of rep1 vs rep2 to see how they fall on a diagonal line.

Is there something special about this experiment?

If everything is correct, and these really are biological replicates, you can use the two lines of code in the error message to continue. And then use nbinomWaldTest() for testing and results() for building the results table.

ADD REPLYlink written 3.7 years ago by Michael Love20k
0
gravatar for ygchai2015
3.7 years ago by
ygchai20150
Korea, Republic Of
ygchai20150 wrote:

I appreciate your comments. I have solved the problem. It is not a problem of DESeq2. I am terribly sorry about this. 

ADD COMMENTlink written 3.7 years ago by ygchai20150
1

thanks for the follow-up. Can you explain what happened? If it happened to one person, chances are it will happen to someone else in the future, and we can ask people to check if they come upon the same issue in the future.

ADD REPLYlink written 3.7 years ago by Michael Love20k
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 2.2.0
Traffic: 233 users visited in the last hour