Determining the Likelihood of Disease Pathogenicity Among Incidentally Identified Genetic Variants in Rare Dilated Cardiomyopathy-Associated Genes

Abstract

Background As utilization of clinical exome sequencing (ES) has expanded, criteria for evaluating the diagnostic weight of incidentally identified variants are critical to guide clinicians and researchers. This is particularly important in genes associated with dilated cardiomyopathy (DCM), which can cause heart failure and sudden death. We sought to compare the frequency and distribution of incidentally identified variants in DCM-associated genes between a clinical referral cohort with those in control and known case cohorts to determine the likelihood of pathogenicity among those undergoing genetic testing for non-DCM indications. Methods and Results A total of 39 rare, non-TTN DCM-associated genes were identified and evaluated from a clinical ES testing referral cohort (n=14 005, Baylor Genetic Laboratories) and compared with a DCM case cohort (n=9442) as well as a control cohort of population variants (n=141 456) derived from the gnomAD database. Variant frequencies in each cohort were compared. Signal-to-noise ratios were calculated comparing the DCM and ES cohort with the gnomAD cohort. The likely pathogenic/pathogenic variant yield in the DCM cohort (8.2%) was significantly higher than in the ES cohort (1.9%). Based on signal-to-noise and correlation analysis, incidental variants found in FLNC, RBM20, MYH6, DSP, ABCC9, JPH2, and NEXN had the greatest chance of being DCM-associated. Conclusions The distribution of pathogenic variants between the ES cohort and the DCM case cohort was gene specific, and variants found in the ES cohort were similar to variants found in the control cohort. Incidentally identified variants in specific genes are more associated with DCM than others.

Keywords: dilated cardiomyopathy; exome sequencing; genetics; incidental finding; secondary finding.

Figure 1. Prevalence of variants in dilated cardiomyopathy (DCM)–associated genes in the exome sequencing (ES) cohort.

A, Schematic representing the derivation of the ES cohort. The clinical exome sequencing (ES) cohort consisted of variants that qualified for clinical reporting and excluded test validation studies and oncological samples. Silent variants and benign variants were excluded. B, Pie chart demonstrating individuals with no variants in non‐TTN DCM‐associated genes (negative, white fill), individuals with variants interpreted as variants of uncertain significance (VUS) (light blue fill) and individuals with variants interpreted as likely pathogenic or pathogenic (LP/P) (dark blue fill). C, Pie chart representing individuals with at least 1 positive variant. D, Bar graph of the ES cohort demonstrating the overall variant yield by gene.

Figure 2. Bar graphs of the pathogenicity‐specific yield by cohort.

A, Graph of the likely pathogenic or pathogenic (LP/P) variant yield for the dilated cardiomyopathy (DCM), exome sequencing (ES), and gnomAD cohorts A). B, Graph of variants of unknown significance (VUS) variant yield for the ES and gnomAD cohorts. To assign pathogenicity to gnomAD variants, ClinVar pathogenicity assignments were referenced. *Represents a significant difference in yield between cohorts (P<0.017).

Figure 3. Bar graphs showing the gene‐specific variant yield among the 3 cohorts.

Panels include likely pathogenic or pathogenic (LP/P) variant yield in the dilated cardiomyopathy (DCM) case cohort (A), variant yield in the exome sequencing (ES) cohort split into LP/P (B) and variants of unknown significance (VUS) (C), and overall variant yield in the gnomAD cohort (D). Error bars represent 95% CIs. *Represents a yield significantly >0 (P<0.0013).

Figure 4. Graphs depicting variant yield by dilated cardiomyopathy (DCM)–associated gene in the DCM and exome sequencing (ES) cohorts, by pathogenicity, with the gnomAD cohort overall yield.

Displayed comparisons include variant yields in the DCM vs ES cohort (likely pathogenic or pathogenic [LP/P]) (A), the DCM vs gnomAD cohort (B), the ES (LP/P) vs gnomAD cohort (C), and the ES (variant of unknown significance [VUS]) vs the gnomAD cohort (D). All variants in the DCM cohort are LP/P. R ² values represent the square of the Pearson correlation coefficient. Each point represents a single gene, labeled when possible.

Figure 5. Graphs depicting variant yield by non‐TTN, dilated cardiomyopathy (DCM)–associated gene in the DCM and exome sequencing (ES) cohorts, by pathogenicity, with the protein size (amino acid length).

Displayed comparisons include variant yields in the DCM cohort (A), the ES cohort (likely pathogenic or pathogenic [LP/P]) (B), the ES cohort (variant of unknown significance [VUS]) (C), and the gnomAD cohort (all variants) (D). All variants in the DCM cohort are LP/P. R ² values represent the square of the Pearson correlation coefficient. Each point represents a single gene, labeled when possible.

Figure 6. Diagram of genes probabilistically associated with dilated cardiomyopathy (DCM).

Genes from the DCM cohort with the highest frequency of variation relative to the gnomAD cohort (signal‐to‐noise ratio) and protein size (left box) and the genes with the highest gene‐specific likely pathogenic or pathogenic (LP/P) variant yields in the exome sequencing (ES cohort) (right box) were compared to identify genes common to all analyses likely to be associated with DCM (center box).