I have a statistics question.
My experimental setup is the following:
Conditions:
cell age: young, old
treatment: control, drug
Different genome readouts (methylation, acetylation, ...) were taken for these conditions. The processed (e.g. CPM normalization) signal for each readout was organized into a matrix with rows equaling genomic location and columns equaling readout type/condition.
Using PCA for each readout I know that old vs. young separates control samples very well with most of the variance explained (~70%) for all readouts.
How can I test whether the drug significantly affects/alters the variance associated with cell age? I.e. say methylation changes in old vs young are reduced stronger by the drug than say acetylation changes?
Thank you!
