Developing new drugs requires substantial financial and time investments, with each candidate undergoing rigorous preclinical and clinical trials. Despite these efforts, many promising candidates fail during clinical trials. A key factor contributing to this may be the reliance on genetically identical animals and monoclonal cell lines in preclinical trials, which fail to represent the diversity of real-world populations. Moreover, preclinical data is often reported in aggregate, obscuring individual-level insights that could be critical for predicting clinical success. In this talk, I will present case studies of how our Standing Variations Modeling technique can address these challenges by: (1) deconstructing aggregate data from preclinical studies to recover individual-level insights ('who’s in'); and (2) extrapolating these findings to a more diverse population by conducting a virtual preclinical trial ('who’s out’).