Fig. 5From: CONFINED: distinguishing biological from technical sources of variation by leveraging multiple methylation datasetsHighlighting treatment effect. We removed from a dataset with simulated treatment effect the components generated by CONFINED. Notably, this simulated treatment effect was not shared across datasets. On the left, PCA performed on the dataset prior to removing the CONFINED components, and on the right the PCA of the dataset after regressing out the CONFINED componentsBack to article page