This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Deseq2 - how to save zero counts on normalization step

Hi all,

I have raw read counts [of rna-seq]. I need to normalize this data and have done it using DESeq2. But, I found that DESeq2 eliminates the read counts of genes when it's equal to 0. I don't want to lose this data as it might affect my result.

Is there any way to do the normalization without eliminating zero read counts?

Thanks in advance.

deseq2 normalization
• 3,463 views • link
written by Sandhiya • Sandhiya

The author suggest removing empty genes are you sure you know better than he? I think empty genes are not removed by default, so can't you just not remove them?

Genes with read count 0 will be ignored when normalizing. They have to be.

• 0 views • link
written by swbarnes2 swbarnes2

@swbarnes2 I don't compete OR get into argument with what authors say.. I'm writing my requirement and how I can find a solution for it.

• 0 views • link
written by Sandhiya • Sandhiya

1 answer

DESeq2 does not remove any genes by default (0 counts or not). Your results will contain all genes, though those that don't pass independent filtering due to low expression will have adjusted p-values set to NA.

Similarly, 0's are not removed from the normalized counts, e.g. min(counts(dds, normalized = TRUE)) will still be 0.

• 0 views • link
written by jared.andrews07 jared.andrews07

Thank you @jared.andrews..

I have solved this ..

• 0 views • link
written by Sandhiya • Sandhiya

Log in to answer this question.

More posts like this