The Problem of Computing FPKM by DESeq2

Question

[gengyang12138]

I have got some different expression gene and their Fold Change, but I want to look over their FPKM in different phrases. So I computed the FPKM of gene by used the command fpkm ( ) of DESeq2. And I find the FPKM of some different expression genes are zero. For example:

The log2FC of gene is not zero, but the FPKM is zero.

gtf = "analyzing data/Ailuropoda_melanoleuca.ailMel1.90.gtf"
txdb = makeTxDbFromGFF(gtf, format = "gtf", circ_seqs = character())
ebg = exonsBy(txdb, by = "gene")
DE_FPKM = DESeqDataSetFromMatrix(countData = count, colData = coldata, design = ~ condition)
rowRanges(DE_FPKM) = GRangesList(ebg)
FPKM = fpkm(DE_FPKM)
write.csv(FPKM, file="FPKM.csv")

Above is my code.

And I want to know where did wrong lead to such a result.

Thank you so much for your suggestions

[WouterDeCoster]

I'd suggest you don't use FPKM, but rather the normalized counts (vst/vsd/rld) generated by DESeq2.

[Kevin Blighe]

I am in complete agreement with Wouter, i.e., not to use FPKM.

In addition, please look up the DESeq2 Tutorial. You should have, as input [to DESeq2], raw counts.