Hi!

I've been performing RNAseq analysis by using svaseq package for batch effect correction. I have identified two slightly correlated known batch effect variables (such as; flowcell, date of data prep.). My question is if I use those two  correlated (r=0.81) known batch effect varaibles in my models, I would overcorrect the data, or since they are slightly correlated I should use both of them and sva would take this into consideration. 

Thank you in advance.

Best,

Burcu