Signal-to-Noise Analysis Can Inform the Likelihood That Incidentally Identified Variants in Sarcomeric Genes Are Associated with Pediatric Cardiomyopathy

Abstract

Background: Hypertrophic cardiomyopathy (HCM) is the most common heritable cardiomyopathy and can predispose individuals to sudden death. Most pediatric HCM patients host a known pathogenic variant in a sarcomeric gene. With the increase in exome sequencing (ES) in clinical settings, incidental variants in HCM-associated genes are being identified more frequently. Diagnostic interpretation of incidental variants is crucial to enhance clinical patient management. We sought to use amino acid-level signal-to-noise (S:N) analysis to establish pathogenic hotspots in sarcomeric HCM-associated genes as well as to refine the 2015 American College of Medical Genetics (ACMG) criteria to predict incidental variant pathogenicity. Methods and Results: Incidental variants in HCM genes (MYBPC3, MYH7, MYL2, MYL3, ACTC1, TPM1, TNNT2, TNNI3, and TNNC1) were obtained from a clinical ES referral database (Baylor Genetics) and compared to rare population variants (gnomAD) and variants from HCM literature cohort studies. A subset of the ES cohort was clinically evaluated at Texas Children’s Hospital. We compared the frequency of ES and HCM variants at specific amino acid locations in coding regions to rare variants (MAF < 0.0001) in gnomAD. S:N ratios were calculated at the gene- and amino acid-level to identify pathogenic hotspots. ES cohort variants were re-classified using ACMG criteria with S:N analysis as a correlate for PM1 criteria, which reduced the burden of variants of uncertain significance. In the clinical validation cohort, the majority of probands with cardiomyopathy or family history hosted likely pathogenic or pathogenic variants. Conclusions: Incidental variants in HCM-associated genes were common among clinical ES referrals, although the majority were not disease-associated. Leveraging amino acid-level S:N as a clinical tool may improve the diagnostic discriminatory ability of ACMG criteria by identifying pathogenic hotspots.

Keywords: exome sequencing; genetic testing; hypertrophic cardiomyopathy; incidentally identified variant; mutation hotspot; secondary finding.

Figure 1

(A), Schematic of the study methodology. (B), Pie chart of all subjects undergoing ES testing divided into individuals with no HCM gene-associated variant (variant-negative; grey) and individuals who hosted a variant of uncertain significance (VUS; white) or likely pathogenic/pathogenic variants (LP/P, red) according to laboratory interpretation at the time of reporting. (C), Pie chart of ES cohort demonstrating variant-positive individuals with a single variant (white) and those with two or more variants (blue).

Figure 2

(A), Bar graph comparing total, radical, and missense variant frequency of HCM-associated variants. (B), Bar graph comparing variant frequency across genes. Comparisons made across a control cohort (gnomAD, white), a cohort of ES referrals (blue), and a cohort of individuals with clinical HCM (green). Error bars denote 95% CI. *, p < 0.05.

Figure 3

(A), Bar graph demonstrating the gene-level signal-to-noise ratios for each HCM-associated gene for the HCM case cohort compared with the gnomAD cohort. (B), Bar graph demonstrating the signal-to-noise ratios for each HCM-associated gene for the ES cohort compared with the gnomAD cohort. Error bars denote 95% CI.

Figure 4

(A), Amino acid-level signal-to-noise analysis of MYH7 variants for both HCM cases (red) and ES-identified variants (blue), compared with variants identified in the gnomAD cohort. (B), Amino acid-level signal-to-noise analysis of MYBPC3 variants for both HCM cases (red) and ES-identified variants (blue), compared with variants identified in the gnomAD cohort. Functional domains of MYH7 and MYBPC3 are depicted. N, N-terminal domain; Cx, immunoglobulin-like domain; FNx, fibronectin domain; highlighted yellow regions, actin binding sites.

Figure 5

(A), Bar graph demonstrating proportion of pathogenic (P), likely pathogenic (LP), and variants of uncertain significance (VUS) using American College of Medical Genetics and Genomics (ACMG) criteria before and after re-assignment with signal-to-noise (S:N). (B), Exome sequencing (ES) cohort was assigned pathogenicity based on 2015 ACMG criteria. A retrospective clinical analysis was performed on those patients seen at Texas Children’s Hospital (TCH) following exclusion of subjects with structural heart disease, mitochondrial disease, or chromosomal abnormalities, or those without echocardiogram. (C), Bar graph showing the percentage of individuals with cardiomyopathy (CM) or first-degree family history (FHx) of cardiomyopathy out of TCH cohort probands hosting VUSs or LP/P variants. (D), Bar graph highlighting the percentage of individuals hosting LP/P variants out of TCH cohort probands with a negative clinical evaluation or cardiomyopathy or first-degree family history of cardiomyopathy. ***, p < 0.001.