Dear all,

My question was partially addressed in similar threads, but I'd like to raise the point one more time.

I have read counts that come form RNAseq data, which was mapped to diploid genome of hybrid yeast (there are 3 biological replicates), thus I have counts for both parental homologs.

    Count matrix looks like:

        Par1_rep1  Par1_rep2   Par1_rep3   Par2_rep1   Par2_rep2   Par2_rep3
gene1       1405         697        1594         992         367        1081
gene2        219         220         259         246         229         272
gene3        896         799         814         498         438         410
gene4         27          59         106          36          62          40
gene5        853        1638        1995         877        2624        2239

What I want is to asses allele-specific expression between two parents using Deseq2. My overall calculations look like

groups<-factor(x=c(rep("par1",3), rep("par2",3)), levels=c("par1","par2"))

colData <- DataFrame(condition=groups)

dds_hybrid <- DESeqDataSetFromMatrix(countData, colData, formula(~condition))

sizeFactors(dds_hybrid) = c(rep(1,6))

dds_hybrid <- DESeq(dds_hybrid)

res_hybrid <- results(dds_hybrid, cooksCutoff=FALSE)

I'd appreciate very much if someone can comment of this and point out if something is wrong.
Especially I am interested whether in this kind of design SizeFactors should be set to 1.

Cheers,

Here's one way to go about looking for ASE with RNA-seq DE tools:

http://rpubs.com/mikelove/ase

This kind of modeling has a pair of counts for each sample: the count for ref and for alt (or you can label them however you like) and looks for genes where the ratio changes across condition.