What's the different between Log2(FPKM) and log2foldchange when we analyse the differential expression?

Question

[Sindy]

A simple question about the differential expression. When I analyse the differential expression and heatmap, I was confused about the Log2(FPKM) and log2foldchange. I only know Log2(FPKM) is a method of normalization, and it can used for building heatmap, but what the function about log2foldchange? Only uesd for knowing which data was up-regulation or down-regulation? but, Log2(FPKM) also can it.

Answer 1

[Benn]

Yeah, you give your own answer already. FPKM values are (normalized) expression values, each sample will have one for each gene. Fold changes are ratio's between two samples (or better groups). So then you get up or down regulation.

[Sindy]

Supposing one data that control：0.04， drought 1h：1.73, drought 6h:4.15. But the heatmap shows that the all the differential datas range from 0 to 10. What's the difinition about up-regulation/down-regulation? Is drought/control >2 consider as up-regulation or the data which greater than 5 ?

[Benn]

You'll need a statistical test to determine if something is up or down regulated. Fold change cut off is not a good threshold for this.

I suggest you use limma or edgeR for this.