DESeq2/Differential expression analysis without replicates
1
0
Entering edit mode
@priyadalvi-14455
Last seen 2.6 years ago
Germany

Hello Everyone,

For exploratory reasons we have RNAseq data without replicates, I totally understand that it is not possible to estimate any sgnificance from this data as there are no replicates, the purpose is only for exploratory reasons. I am new to bioinformatics and would appreciate any help and sugggestions.

I would like to perform differential expression analysis to understand which genes are majorly affected (Foldchange) in the "control" versus "treatment". I have read the previous posts about this kind of analysis but without any luck.

Previous posts I have read and followed:

http://seqanswers.com/forums/showthread.php?t=31036

Deseq2 for RNAseq experiments without replicates

DESeq2, no replicate

Below is an example of the code I am running and the errors I encounter:

library(DESeq2)

countdata<-read.table("T1.txt", header=TRUE, row.names=1, check.names = F)

class(countdata)

countdata<-as.matrix(countdata)

class(countdata)

condition<-factor(c("Control,"Treatment"))

(coldata<-data.frame(row.names=colnames(countdata), condition))

coldata

dds<-DESeqDataSetFromMatrix(countData =  countdata, colData = coldata, design=~condition)

dds <- DESeq(dds)

# Error :
estimating size factors
estimating dispersions
Failure in checkForExperimentalReplicates(object, modelMatrix) : 
The design matrix has the same number of samples and coefficients to fit, so estimation of dispersion is not possible. Treating samples as replicates was deprecated in v1.20 and no longer supported since v1.22.

DESeq(dds)

# Error :
estimating size factors
estimating dispersions
Fehler in checkForExperimentalReplicates(object, modelMatrix) : 
The design matrix has the same number of samples and coefficients to fit, so estimation of dispersion is not possible. Treating samples as replicates was deprecated in v1.20 and no longer supported since v1.22.

res <- results(dds)

# Error:
Fehler in results(dds) : 
couldn't find results. you should first run DESeq()

#

Any help and pointers in this direction would really be helpful for me. If anyone could share the code to use with the latest version of DESeq2 for this kind of analysis would be great. Thank you very much in advance :)

Cheers, PD

DESeq2 DEG • 5.6k views
ADD COMMENT
0
Entering edit mode
@mikelove
Last seen 2 hours ago
United States

You can calculate vst() on a dds with a design of ~1 and then subtract the two columns to form a LFC.

We don't offer any inference on design without replicates through DESeq().

ADD COMMENT
0
Entering edit mode

Thanks for your reply Michael. I will check if this could help me. If not DESeq, is there any other Package you are aware of hat could be more suitable for me to use? Thank you.

ADD REPLY
0
Entering edit mode

I don't follow, I told you how to compute a stabilized LFC from one replicate data using DESeq2.

ADD REPLY
0
Entering edit mode

Hi Michael, thank you very much for your reply. I followed your suggestion have found my way to get normalised counts and calculate LFC. Thanks again :)

ADD REPLY
0
Entering edit mode

You can use NOISeq, which can perform differential expression analysis without replicates.

NOISeq

Refer 5.1.2 NOISeq-sim: no replicates available. Note this method will simulate replicate with normal distribution, but still better than others.

ADD REPLY

Login before adding your answer.

Traffic: 1109 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6