after I use sva package to remove batch effect,can I use DESeq2 to make DE analysis?

Question

[Ada]

hello,here my question. If I use sva packages to remove the batch effect,which mean I need to normalize the read counts.But I want to make DE analysis with DESeq2 ,and that need raw read counts without normalization.and here the conflict is that I need to remove batch effect before differential expression analysis.How can I solve this problem ? I need your help .

Answer 1

[Devon Ryan]

Please have a read through this bioconductor workflow. It shows your exact use case of combining SVA with DESeq2. The short answer from that is that you should incorporate the output from SVA in the model matrix and leave the counts untouched.

[Ada]

Thank you very much!