Open cansavvy opened 5 years ago
We'll use PCA component scatterplot before batch correction and after batch correction as a control of sorts.
Going to re-tackle this one. May look at using removeBatchEffect
from limma
instead of ComBat
? Will see what works.
Where are we at with a batch effect example? Do we still want to try to add a microarray and/or RNA-seq example? If yes, then do we want to do this before going live
or should this continue to be in the Icebox
?
Or is batch correction something that requires too much nuance to be responsibly taught through a tutorial?
This is an advanced usage situation in my opinion.
Would be good to show examples of how to correct for batch effect. For this example, we would use refinebio data that is aggregated by species. Presumably include some sort of initial litmus test of sorts that if followed, would help identify whether batch correction is needed in the first place.