The past decade has brought a wealth of information about the genetic and molecular characteristics of medulloblastoma, a cancer of the cerebellum. This should enable more targeted therapy better adapted to the diversity of the disease. However, barriers remain to translating this knowledge into improved therapeutic outcomes. Bridging the gap between genomics and treatment will require realistic preclinical models and integration of functional and genomic data. To this end, we performed in vitro screening of 7729 drugs on 20 patient-derived xenografts (PDXs) from the three most common medulloblastoma subgroups. One drug, actinomycin D, was shown to be effective in vivo in PDXs of Group 3 medulloblastoma, the molecular subgroup with the worst outcomes. Screening results were analyzed alongside mutational, transcriptional, and epigenetic data. I illustrate how this data could be used to provide personalized therapy recommendations and to shed light on molecular mechanisms underlying differences in drug response.