One assumption is, that the majority of genes are not differentially expressed.
• 0 views
•
link
written
by
caggtaagtat
Depends if you filter out 20 or 20000 genes. Please add details.
Of course, I know I can filter, people filter for example low expressing genes, etc. I was hoping for guidance and insights on the statistical assumptions to better understand how I can partition my data.
Why do you want to filter? Default settings are fine in most cases. How many genes, any specific kind of genes, or just lowly expressed?
Unfortunately, I cannot add more details. This information should be enough to answer my original question: Does DESeq2 expect a minimum number of genes?
All statistical testing your doing depends on the size of elements, so, if you reduce the number of elements, your p-value distribution changes.
Sure. This doesn't answer my question.
DESeq2 don't expect a minimal number, but you are modifying your p-values, so you get less DEGs
Answered by the DESeq2 author: https://support.bioconductor.org/p/131401/