Question

how to use Normalized gene expression data in Deseq2?

0

Entering edit mode

mheydarpour ▴ 10

@mheydarpour-9430

Last seen 5.2 years ago

I have normalized gene-expression data for about 23000 genes. The values are not integers.

So, when I run the dds <- on Deseq2, I got the following error:

"some values in assay are not integers"

Is there any commands in Deseq2 to fix this error? I mean how I can use normalized data directly into Deseq2?

Thanks

Deseq2 • 3.7k views

ADD COMMENT • link updated 8.3 years ago by Gordon Smyth 50k • written 8.3 years ago by mheydarpour ▴ 10

0

Entering edit mode

It would help to give more background about what your expression data represent. Are they from RNA-seq or from some other technology? By what method were the normalized expression values computed?

ADD REPLY • link 8.3 years ago Gordon Smyth 50k

score 0 · Answer 1 · 2015-12-28

0

Entering edit mode

Ryan C. Thompson ★ 7.9k

@ryan-c-thompson-5618

Last seen 8 months ago

Scripps Research, La Jolla, CA

You can't. The statistical model is fundamentally designed for, and based on the assumption of, input data on a raw count scale, and forcing it to analyze non-count data as counts will give you incorrect results. You either need to obtain the raw counts, or use another statistical method that does not rely on having them.

ADD COMMENT • link 8.3 years ago Ryan C. Thompson ★ 7.9k

score 0 · Answer 2 · 2015-12-28

0

Entering edit mode

Michael Love 41k

@mikelove

Last seen 1 day ago

United States

See our vignette : vignette("DESeq2") ... for details on the statistical model and on how the input is expected to be unnormalized counts. You can also find a link to the DESeq2 paper for more details.

ADD COMMENT • link 8.3 years ago Michael Love 41k