Polygenic Background and Penetrance of Pathogenic Variants in Hypertrophic and Dilated Cardiomyopathies

Figures

• 

  Figure 1:. Sample filtering and workflow. From an original sample of unrelated participants, the subset of patients with echocardiogram data was identified. Individuals were separately classified for HCM or DCM case and non-case status, excluding individuals with conflicting diagnosis codes, heart failure, or an HCM phenocopy such as amyloidosis. After validating case definitions with echocardiogram data, associations of genetics with clinical and echocardiogram data were separately tested.

• 

  Figure 2:. Effects of polygenic and monogenic risk on HCM and DCM probability. Logistic regression models including include polygenic risk, monogenic risk, a polygenic:monogenic interaction term, sex, and age at time of biobank enrollment were used for marginal prediction of clinical cardiomyopathy outcomes. A) Prediction of DCM in carriers of a DCM pathogenic variant, using a DCM (left) and HCM (right) PGS. B) Prediction of HCM in carriers of a HCM pathogenic variant, using a DCM (left) and HCM (right) PGS.

• 

  Figure 3:. Contributions of monogenic and polygenic risk terms to A) DCM and B) HCM model discrimination. HCM_P and DCM_P indicate the presence of a pathogenic variant in an HCM and DCM gene respectively.

• 

  Figure 4:. Effects of polygenic and monogenic risk on echocardiogram measurements. Linear regression models including polygenic risk, monogenic risk, a polygenic:monogenic interaction term, sex, and age at time of echocardiogram were used for marginal prediction of echocardiogram measurements. A) Predicted Minimum EF and B) Predicted Maximal LVIDd stratified by DCM pathogenic variant status using DCM (left panel) and HCM (right panel) PGSs. C) Predicted Maximum EF and D) Predicted Maximal IVS stratified by HCM pathogenic variant status using DCM (left panel) and HCM (right panel) PGSs.