The researchers used machine learning to build predictive models for individuals' future risks of nine common diseases, based on their genetic information and the biomarkers that they measured.