Question

Export post-regressed counts with DESeq2

0

Entering edit mode

pchiang5 • 0

@pchiang5-9743

Last seen 8.8 years ago

Hi folks,

Currently I am able to extract a table of counts normalized by size factors:
normalized_counts = counts(dds, normalized=TRUE)

However, these counts are still before fitting in the analysis.

Is it possible to obtain the post-fitting counts from DESeq2 analysis?

Thanks

deseq2 • 1.2k views

ADD COMMENT • link updated 8.8 years ago by Michael Love 43k • written 8.8 years ago by pchiang5 • 0

score 2 · Answer 1 · 2016-02-18

2

Entering edit mode

Michael Love 43k

@mikelove

Last seen 11 hours ago

United States

DESeq2 models the raw counts, using the size factors on the right hand side of the generalized linear model. So there are no post-fitting counts, as opposed to pre-fitting counts.

ADD COMMENT • link 8.8 years ago Michael Love 43k

0

Entering edit mode

If you want to visualize the expression across samples, and perhaps remove fixed effects associated with batches, you could use the VST or rlog transformation (see vignette), then apply limma's removeBatchEffect() to the transformed counts (accessible via assay(object) following a transformation).