Below are distributions of the number of associations per drug for the L1000 and CREEDS datasets. The number of drug-gene associations refers to the number of genes that were significantly affected by the drug perturbation.