**0**wrote:

Hi all experts,

After doing DE analysis by edgeR, we can find the number of up and down-expressed genes by something like, "summary(de <- decideTestsDGE(lrt5))". I would highly appreciate if you could please help me out to figure out the following issues, which I cannot completely understand with reading and browsing net.

**1)** Please let me know how fold change (FC) is calculated during GLM method, it's based on cpm? say we compare A and B conditions, fold change of DE genes are calculated as cpmA / cpmB or there is another story? if your response is positive, why the software doesn't use RPKM(FPKM) that take into account the gene length? is there any priority for using CPM in relative to FPKM?

**2)** Please kindly tell me how we can export DE genes with all related information, like logFC, logCPM, P-value and FDR from edgeR software?

**3)** As I found on net, we can convert cpm to rpkm via this formula: RPKM = 2^(logCPM-log2(geneLength)). Please let me know if the same formula can be ued for FPKM?

**4)** If I correctly understood from edgeR manual, drawing heatmap by logCPM is more desirable than logFPKM? if it's right, please kindly tell me why there is such preference?

Thank you in advance for all your clarifications.

**7.3k**• written 2.7 years ago by Sara •

**0**