Biomarkers that help patients to choose the best treatment option, play a key role in drug development and precision medicine. The talk introduced the Virtual Twin Method by Foster et al. (2011). It is a multivariate, tree-based method that identifies patients who may benefit from a certain treatment. The close relationship to prognostic and predictive biomarkers was illustrated, and the advantages and disadvantages compared to univariate methods were highlighted. This internal Staburo training was a great opportunity to share knowledge within our team and to start a very fruitful discussion about the Virtual Twin Method!