Hi,

I have RNAseq data from two conditions in two different genotypes. I have put the same design as explained in mannual of ?results.

I am interested to know the effect of condition between two different genotype.

Below is the code that i followed

> condition <- factor(c(rep("Uns",4),rep("treat",4)),levels = c("Uns","treat"))
> genotype <- factor(c(rep("WT",2),rep("KO",2),rep("WT",2),rep("KO",2)),levels = c("WT","KO"))
> coldata <- data.frame(condition,genotype)
> coldata
 condition genotype

1       Uns       WT
2       Uns       WT
3       Uns       KO
4       Uns       KO
5     treat       WT
6     treat       WT
7     treat       KO
8     treat       KO

> design <- ~ genotype + condition + genotype:condition
> dds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=design)
> dds <- DESeq(dds)
> res <- results(dds, name="genotypeKO.conditiontreat")
> resdata <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by="row.names", sort=FALSE)

If I look at these two gene, based on "log2FoldChange" value, both seems to be downregulated, but if look at count normalized value in WT and KO, the value is quite opposite. Why is this so, am I missing something here.  could someone please explain this analysis.

Thanks in advance !

You are using an interaction term in the design, and testing the interaction term. This is not a main effect. Please see the diagram in the vignette section on interactions.