Question: DEseq2 replicate not normalizing
gravatar for kjh318
6 months ago by
kjh31810 wrote:

Hi there, This issue has been posted about before but I couldn't find a definitive answer. I noticed that the DEseq2 normalization count table has one replicate that is not normalized. The number of reads in this table is exactly the same number as the htseq-count file for that replicate. I first reran DEseq2 with the same datasets, but still that replicate was not normalizing. Then I went back to the beginning; aligned that data set, did htseq-count, and used this new file in the DEseq2 run. But still, that replicate is not normalizing. I made sure that when redoing all of this, I was using the exact same settings as were used for the samples which were normalizing correctly. I'd really appreciate some more guidance on this! Thanks!

ADD COMMENTlink modified 6 months ago by Jennifer Hillman Jackson25k • written 6 months ago by kjh31810
gravatar for Jennifer Hillman Jackson
6 months ago by
United States
Jennifer Hillman Jackson25k wrote:


I didn't find a Normalized Counts file with unnormalized values in your most current active history at Galaxy Main Ht-seq reports whole numbers, the normalized counts are not. So, I don't think this is a Galaxy tool wrapper problem/bug.

The input content, factor matrix structure, and settings used all make a difference with how normalization is applied (and with how fold change is calculated).

Bioconductor help for DeSeq2:

If that does not clear things up, ask a question at their support site (or search first, then ask if the question is novel):

Thanks, Jen, Galaxy team

ADD COMMENTlink modified 6 months ago • written 6 months ago by Jennifer Hillman Jackson25k

Hi Jen, That is exactly what I'm seeing: 5 of the 6 replicates have values with multiple decimal points but one is just the whole numbers same as the htseq-count file. The normalized counts files in the Py Comparative data set should show this clearly, the SG-2 replicate is not normalizing. I tried reuploading the problem data set and even re-ran all six replicates with a new reference genome and transcriptome. And even then, this one replicate remains unnormalized. I checked the htseq-count file and there seems to be nothing wrong with that file and is similar to the other htseq-count files of different replicates. I also re-uploaded the tabular htseq-count file for that replicate and tried DEseq2 again, with no avail. All of these replicates were run with the exact same settings so I don't understand why 5 would normalize correctly but the other wouldn't. Could it be an issue with the raw fasta file from that run? Any additional guidance would be appreciated. Thanks

ADD REPLYlink written 6 months ago by kjh31810
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 169 users visited in the last hour