Interpretation of rare sequence variants is a key challenge in clinical genetic testing. In the absence of a definitive model to ascertain variant pathogenicity interpretation is usually conducted by combining evidence from multiple sources via heuristic rules and points-based systems. In this paper, we explore a fundamentally different modeling approach – one based on probabilistic graphical models. We present initial attempts at graphical modeling of the variant interpretation task, highlighting the benefits such as transparency of mod- eling assumptions, explainability, sensitivity analysis, etc. while also describing challenges that are to be overcome.