Fig. 1From: Directed acyclic graphs and causal thinking in clinical risk prediction modelingExample of the Markov Blanket (in black) of outcome Y in a simple Directed Acyclic Graph (DAG) with many nodesBack to article page