Fig. 3From: A new approach to identifying patients with elevated risk for Fabry disease using a machine learning algorithmSelected phenotypic features and relative prevalence (portion of patients with evidence of feature) in risk strata, defined following rank-ordering of patients by predicted Fabry disease risk. Darker coloring indicates these features’ increased prevalence in correspondence with increasing risk of Fabry disease. This set of features is a small sample of the hundreds of signals drawn from available data that drove the tool’s analytic performanceBack to article page