Unlog transformed data in DESeq2
1
0
Entering edit mode
@mohammedtoufiq91-17679
Last seen 16 minutes ago
United States

Hi,

I am using DESeq2 for analyzing Illumina RNASeq datasets. I follow the below steps;

  1. Derived raw counts (from featureCounts) > Imported counts to DESeq2
  2. Normalised the counts via an estimation of size factors (counts(dds, normalized = TRUE))
  3. Transformed the data for downstream applications via variance stabilisation (vst) or regularised log (rlog)
vsd <- vst(dds, blind=FALSE)
rld <- rlog(dds, blind=FALSE)
ntd <- normTransform(dds). # this gives log2(n + 1)

My question is if I would like to reverse transform the data in the unlog form - for ntd I would use; dat_unlog = 2^(ntd - 1)

However, If I am interested to unlog the data - from vst or rlog Is it right to use or are there any additional parameters or arguments?

dat_unlog = 2^vst

(OR)

dat_unlog = 2^rlog

Thank you,

Toufiq

RNASeq DESeq2 vst rlog Normalization • 1.8k views
ADD COMMENT
2
Entering edit mode
@mikelove
Last seen 1 hour ago
United States

All three of these give log2 transform so yes 2^x puts it back on the original scale (but with the sequencing depth variation removed).

Just as a note, don't provide those back-transformed counts to DESeq().

ADD COMMENT
1
Entering edit mode

Dear Michael Love noted and thank you very much for the prompt response. This was helpful.

ADD REPLY
0
Entering edit mode

Dear Michael Love, I want to pass RUVIII corrected nanostring data to Deseq2. But it seems that the output (batch normalize count) I am getting from RUVIII is log-transformed. Do you have any suggestions, since we can't pass back-transformed counts to Deseq2 (if I am right)? Thanks. Also wondering if this count is suitable input for limma without any further modification Gordon Smyth thanks.

ADD REPLY
1
Entering edit mode

The RUVIII people themselves use limma ( https://doi.org/10.1093/nar/gkz433 ). But please, post your question as a question. That would be better than commenting on a comment from an unrelated question from two years ago.

ADD REPLY

Login before adding your answer.

Traffic: 767 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6