I will discuss how to leverage large-scale EHR data linked with tumor mutation profiles to emulate cancer clinical trials in silico. This approach helps us to design real trials, making them more incluse and efficient (Liu et al. Nature 2021). It also helps to identify a large number of predictive biomarkers, i.e. mutations that impact the survival of patients who are taking specific chemo, targeted or immuno therapies (Liu et al. Nature Medicine 2022). These mutation-drug interactions are critical for treatment planning and precision oncology.