This is a test version of Biostars. For the public version, visit https://www.biostars.org.
DESeq2: How to interpret numbers for independent filtering?

Hi,

I am doing differential expression using DESeq2 and trying to learn more about independent filtering. I used the function metadata(res)$filterThreshold and got the following numbers but don't quite understand them.

> 66.02757% 
> 6.589014

I understand DESeq2 filters out genes of a mean normalised count below the threshold. In this case, does this mean that DESeq2 has filtered out genes with a mean count of 6.589? And based on this counts threshold, 66% of genes were filtered out of the analysis?

If so, is the mean count threshold abnormally low? Or is this quite normal?

Thank you!

r rna-seq
• 2,423 views • link
updated by rohitsatyam102 rohitsatyam102 • written by nattzy94 • nattzy94

So, this means that genes with basemean < 6.589014 have been filtered and the filtered genes represent 66.02757% of all tested genes!

This is a lot of genes whose P values are being set to NA.

Now you might ask what will you do about it.

  1. First it would be interesting to see if you prefiltered the genes with zero counts before running DESeq2. If you didn't I would recomment to use rowSums function to filter unexpressed genes.
  2. To get a feel of how many genes have basemean==0 is also a way to find out if the majority of genes whose padj value have been turned to NA are the one's with basemean==0 or not.
  3. Try to set independent filtering parameter to FALSE and then see how many genes have NA padj values.
• 0 views • link
written by rohitsatyam102 rohitsatyam102

0 answers

No answers yet.

Log in to answer this question.

More posts like this