Question

very low number in normalized count but give very high in log2foldchange

0

Entering edit mode

naktang1 • 0

@naktang1-8825

Last seen 4.7 years ago

Thailand

I have one gene that have very high in log2foldchange.

gene	baseMean	log2FoldChange	lfcSE	stat	pvalue	padj
geneA	50.46458	41.13637	3.733037	11.01955	3.08E-28	1.19E-23

so the foldchange is 41 while the basemean is 50.46

However, when i check the normalized count and sizefactor

counts(dds["gene.118755",],normalized=T)
              D13_ABC_C1_count  D13_ABC_C2_count   D13_ABC_H1_count    D13_ABC_H2_count
gene.118755         0                 0                 0             1.129077

D13_ABC_C1_count D13_ABC_C2_count D13_ABC_H1_count D13_ABC_H2_count

sizeFactors(dds) 0.7990090 0.8081792 0.5362605 2.6570366

My question is why this gene has a very small in normalized count but still give a very high number in folchange of this gene

I compare between D13_ABC_H vs D13_ABC_C condition.

deseq2 • 811 views

ADD COMMENT • link 6.2 years ago naktang1 • 0

0

Entering edit mode

thank you for you reply, However what threshold level for LFC that you will recommend to use lfcshrink function

ADD REPLY • link 6.2 years ago naktang1 • 0

0

Entering edit mode

We can't provide that because it is specific to your experiment. I don't know what is a biologically meaningful fold change for you samples. I'd recommend coming up with this through discussion with other biologists and investigators (if you are collaborating).

ADD REPLY • link 6.2 years ago Michael Love 41k

score 1 · Answer 1 · 2018-02-28

1

Entering edit mode

Michael Love 41k

@mikelove

Last seen 2 hours ago

United States

The LFC is not defined because you have positive vs zeros. You should use lfcShrink to provide accurate LFC estimates.

ADD COMMENT • link 6.2 years ago Michael Love 41k