DESeq without replicate nbinomTest method

Question

[Bioblazer]

I have 6 samples of RNA-seq files without any technical replicates (Saprophyte_30,Saprophyte_37,Healed_30,Healed_37,surgery_30,surgery_37). These are A.flavus (fungus)samples which were grown at 30 and 37 deg C of saprophytic condition and its clinical isolates. I am following "Count-based differential expression analysis of RNA sequencing data using R and Bioconductor" protocol paper for my data analysis.

I got the error when I executed following command in DESeq

cds=estimateDispersions(cds,method="blind")
res=nbinomTest(cds,"Saphrophyte _30","Saphrophyte _37")

Error in if (dispTable(cds)[condA] == "blind" || dispTable(cds)[condB] == : missing value where TRUE/FALSE needed

Can anyone tell me how to solve this?

[EVR]

Maapu,

Try using GFOLD for samples without replicates. It is better than DESeq.

Answer 1

[mark.ziemann]

Try the following:

cds <-estimateDispersions(cds,method="blind",sharingMode="fit-only",fitType="local") res <- nbinomTest( cds, "Saphrophyte _30", "Saphrophyte _37")

[anne.leinweber]

I have the same problem (two treatments, only one replicate each). I use the code that mark.ziemann wrote, however the same error still occurs. I don't understand why it comes up. What is happening in the background and how can i change my dataset to make the nbinomTest function work? Cheers, Anne

[mark.ziemann]

Hi Anne, maybe open up a new biostars ticket with some additional info like error messages and we can help better.

[anne.leinweber]

Hey Mark! Thank you for replying to my comment.

You can find my problem here: C: error after DESeq function nbinomTest with only one replicate

It is really the same error, even after using your code. But I want to compare two samples without replication.