DESeq2 log2 fold change

Question

[Morris_Chair]

Hello guys,

I'm doing expression profiling analysis and I know that the log2fc indicates how much changes the expression of a condition compared to another. If I have two condition where one is treated and one is untreated, and I set as reference level the treated contitions, the output of DESeq2 refers how many folds the treated contition changes over the untreated, is it correct? I have this doubt ..

Thank you

Answer 1

[Kevin Blighe]

If you set treated as the reference level, as you have done, then a positive log (base 2) fold change means higher expression in untreated. That is, the reference level becomes the denominator in the fold change calculation - everything else is the numerator.

You can easily verify this via a box and whisker plot where you stratify expression of a gene across your 2 conditions, e.g.:

boxplot(condition ~ gene, data = MyData)

[Morris_Chair]

Thank you for the clarifying Kevin