Uneven number of control/treatment replicates

Question

[hac141]

I wonder if anyone has experience doing an analysis with 3 biological replicates for the treatment but only 2 for the control? I am almost certain I screwed up one of my control samples during the libraries construction and now PCA analysis shows that one of my control libraries clusters with the treatment libraries. Obviously DESEq is giving me basically no significant DE, while removal of the "bad" library gives me a result that makes sense.

[GZ1995]

Usually no need to worry about uneven group sizes.