comparision pairwise samples with DESeq2

Question

[lkianmehr]

Hello

I need your help to resolve this problem with DESeq2 because It's my first experience with DESeq2! I have read other related posts, but I can't understand how can do it for my samples. I have 8 samples (count file) and want to compare them pairwise! I need to compare :

1-F0N D-RNA3 / to F0F-D-RNA5

2- F0N –RNA3 / to F0F- RNA5

3- F0N D-RNA3 / to F1F-D-RNA8

4- F0N RNA3 / to F1F- RNA8

and, I am starting with these codes:

options(stringsAsFactors =F)
setwd("/media/leila/LaCie/Data_colorado/Count/Series1/MM/")
dds <- estimateSizeFactors(dds)

Error in estimateSizeFactors(dds) : object 'dds' not found

results(dds, contrast = c("F0N_D_RNA_3", "F0F_D_RNA_5"))

Error in is(object, "DESeqDataSet") : object 'dds' not found

I got some errors, Does anybody know how to resolve it?

thanks in advance

[ATpoint]

This tutorial on DESeq2 should answer your question. Focus on the part about contrasts in the section "Variations to the standard workflow".

[h.mon]

I have 8 samples (count file) and want to compare them pairwise!

Does this means you don't have biological replicates? If you don't, DESeq2 is not the appropriate package for the analysis. More importantly, having no biological replicates is a really bad experimental design, and your results will be full of false positives and false negatives.

Anyway, you want to use the contrast argument of results(), something akin to:

results(dds, contrast=c("group","F0N-D-RNA3 ","F0F-D-RNA5"))

[lkianmehr]

I got this error: Error in is(object, "DESeqDataSet") : object 'dds' not found

[h.mon]

1) Please read carefully the link posted by ATpoint above, and try to follow it:

http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html

2) Edit your original question and post the code you are using.

edit: you didn't answer if you do or don't have biological replicates.

[lkianmehr]

Sorry, F0F, F0N, and F1F samples are same but the type of sequence is different, one of them is Free-RNA and another is DNA-associated RNA (from Mouse). furthermore, I have 4 samples like that from the human. at first, I am going to compare Mouse samples (with the above order), then compare them versus human samples. but I don't have one more of each one! I'd appreciate if explain me details about the related codes!

thank you very much

[h.mon]

How did you obtain the counts? How are they organized? You have to read the counts into R, then use one of the appropriate functions to create a DESeqDataSet object.

Please do read the DESeq2 vignette, it is all covered there.

http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html

[lkianmehr]

I did it by HTSeq-count analysis. what do you mean about an organization? alright. just because I am new with R and don't have enough time to be expert wanted to help me.

[h.mon]

How are they organized?

What I meant by that is: several files with counts, or just one file? Scattered across several folders, or located in one folder? And so on.

There is a section called htseq-count input on the DESeq2 vignette. Try to follow it and report back if you encounter any errors.

[lkianmehr]

I have 2 series samples. series 1 include above data from Normal and high-fat diet samples and Series 2 include mutant mouse data with Normal and high-fat diet. series 1 data are in 2 folders (Mouse and Human) and Series 2 are in another folder. at first, I want to compare each one pairwise and then together. for example: in series1:

1- F0N D-RNA3 / to F0F-D-RNA5 and ...

2- then all free-RNA versus D-RNA .

in series2:

ND (2 samples) versus HD (2samples)

NF (2samples) versus HF (2samples)

finally:

F0N from series1 vs ND (2samples) from Series2 and ...