Common genetic modifiers influence cardiomyopathy susceptibility among the carriers of rare pathogenic variants

Figure 1. Identification of known and predicted rare pathogenic variants for HCM and DCM in the UK Biobank exome data (A) Computational workflow used to identify known pathogenic/likely pathogenic (PLP) variants in ClinVar as well as predicted loss-of-function (pLoF) and predicted damaging missense (pDM) variants. (B) Number of known (green) and predicted (orange) pathogenic variants. (C) Number of carriers of known (green) and predicted (orange) pathogenic variants. AM, AlphaMissense prediction score.

Figure 2. Association of known and predicted rare pathogenic variant classes with HCM and DCM disease risk (A and B) Cumulative incidence of HCM or DCM (%) is plotted using Kaplan-Meier (KM) curves. (C and D) Forest plots illustrate the association of variant classes with HCM and DCM risk using CoxPH models. Error bars denote 95% confidence interval. Noncarriers (nc), participants who do not carry known or predicted rare pathogenic variants; PLP, carriers of known rare pathogenic variants in ClinVar; pLoF, carriers of predicted loss-of-function variants; pDM, carriers of predicted damaging missense variants; nvariants, number of variants; ncarrier, number of carriers; ndisease, number of carriers with disease; HR, hazard ratio; SE, standard error; p, p value.

Figure 3. Association of known and predicted rare pathogenic variant classes with HCM and DCM age of diagnosis among disease cases Violin and boxplots display the distribution of age of diagnosis among HCM (A) and DCM (B) carrier groups (top). Forest plots show the association of rare pathogenic variant classes with age of diagnosis for HCM (A) and DCM (B) using linear regression model (bottom). Error bars denote 95% confidence interval. Noncarriers, participants who do not carry known or predicted rare pathogenic variants; PLP, carriers of known rare pathogenic variants in ClinVar; pLoF, carriers of predicted loss-of-function variants; pDM, carriers of predicted damaging missense variants; nvariants, number of variants; ncarrier, number of carriers; β, estimated effect; SE, standard error; p, p value.

Figure 4. Association of known and predicted rare pathogenic variant classes with cardiac phenotypes representative of HCM and DCM disease progression among disease cases (A) Violin plots show the distribution of mean left ventricular myocardial wall thickness (LVWT) per HCM disease status and rare pathogenic variant class. (B) Violin plots show the distribution of left ventricular ejection fraction (LVEF) per DCM disease status and DCM rare pathogenic variant class. (C) Association of rare pathogenic HCM variants with mean LVWT using linear regression. (D) Association of rare pathogenic DCM variants with LVEF using linear regression. Error bars denote 95% confidence interval. Noncarriers, participants who do not carry known or predicted rare pathogenic variants; PLP, carriers of known rare pathogenic variants in ClinVar; pLoF, carriers of predicted loss-of-function variants; pDM, carriers of predicted damaging missense variants; nvariants, number of variants; ncarrier, number of carriers; β, estimated effect; SE, standard error; p, p value.

Figure 5. HCM PRS is associated with disease risk among the carriers of known and predicted rare pathogenic variants (A) HCM PRS distribution is shifted upward in the subjects with HCM compared to control subjects among the rare pathogenic carriers (S3). (B) Cumulative incidence of HCM (%) among the rare pathogenic variant carriers (S3) and noncarriers stratified by the HCM PRS is plotted using KM curves with the number at risk and number of events at each time point summarized in the table on the bottom. (C) Forest plot illustrates the association of PRS with HCM risk using CoxPH models among the noncarrier and carrier sets (S1, S2, and S3). (D) Forest plot shows the effect of both the rare pathogenic variant and common variant PRSs on HCM risk, with the noncarriers with intermediate PRSs as the reference. Error bars denote 95% confidence interval. PRS bins: low (bottom 20%), intermediate (middle 60%), and high (top 20%). n, number of individuals in carrier sets; ndisease, number of HCM cases in the carrier sets; HR, hazard ratio; SE, standard error; p, p value.

Figure 6. DCM PRS is associated with disease risk among the carriers of known and predicted rare pathogenic variants (A) DCM PRS distribution is right shifted in the subjects with DCM compared to control subjects among the rare pathogenic variant carriers (S3). (B) Cumulative incidence of DCM (%) among the rare pathogenic variant carriers (S3) and noncarriers stratified by DCM PRS is plotted using KM curves with the number at risk and number of events at each time point summarized in the table on the bottom. (C) Forest plot illustrates the association of DCM PRS with DCM risk using CoxPH models among the noncarrier and carrier sets (S1, S2, and S3). (D) Forest plot demonstrates the effect of both the rare pathogenic variant and common variant PRSs on DCM risk, with the noncarriers with intermediate PRSs as the reference. Error bars denote 95% confidence interval. PRS bins: low (bottom 20%), intermediate (middle 60%), and high (top 80%). n, number of individuals in carrier sets; ndisease, number of individuals with DCM in the carrier sets; HR, hazard ratio; SE, standard error; p, p value.