Nested analysis or Collapsed-replicate analysis?
1
0
Entering edit mode
@b295d7f1
Last seen 2 days ago
Italy

Hi,

I am performing a DE analysis using Deseq2. For my experimental settings, I would like to compare leukodystrophy affected patient against control patient. For each condition, I have three patients and two replicates from the same patient, with a total of 6 replicates for each condition (see Figure attached). enter image description here

I am attaching also the sample distribution with PCA plots with condition variable and the information of patient

enter image description here

enter image description here

I was wondering if for this experiment it is recommended to perform a nested analysis or a “collapsed- replicates” analysis.

Best regards,

Giulia Lopatriello

RNASeq DESeq2 • 133 views
ADD COMMENT
2
Entering edit mode
@mikelove
Last seen 2 days ago
United States

I'd use limma with duplicateCorrelation. These do not look like "technical replicates" as we use the term in the vignette (additional sequencing of prepared libraries).

ADD COMMENT
0
Entering edit mode

Dear Michael,

I have run the analysis with limma and I have obtained no significant gene whereas in DESeq2 I have found approximately 100 DE genes.

Given these results, would you strongly recommend to use limma analysis with this type of experiment or could we use DESeq2?

Best regards,

Giulia L.

ADD REPLY
1
Entering edit mode

A key aspect of correct analysis is identifying the unit of analysis. If you want to compare across patients, you can't add replicates per patient without letting the model know that these are not i.i.d. observations. So yes, you will obtain FP by ignoring the correlations of these replicates, and you need to use duplicateCorrelation here.

ADD REPLY

Login before adding your answer.

Traffic: 552 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