Amino acid-level signal-to-noise analysis of incidentally identified variants in genes associated with long QT syndrome during pediatric whole exome sequencing reflects background genetic noise

Abstract

Background: Due to rapid expansion of clinical genetic testing, an increasing number of genetic variants of undetermined significance and unclear diagnostic value are being identified in children. Variants found in genes associated with heritable channelopathies, such as long QT syndrome (LQTS), are particularly difficult to interpret given the risk of sudden cardiac death associated with pathologic mutations.

Objective: The purpose of this study was to determine whether variants in LQTS-associated genes from whole exome sequencing (WES) represent disease-associated biomarkers or background genetic "noise."

Methods: WES variants from Baylor Genetics Laboratories were obtained for 17 LQTS-associated genes. Rare variants from healthy controls were obtained from the GnomAD database. LQTS case variants were extracted from the literature. Amino acid-level mapping and signal-to-noise calculations were conducted. Clinical history and diagnostic studies were analyzed for WES subjects evaluated at our institution.

Results: Variants in LQTS case-associated genes were present in 38.3% of 7244 WES probands. There was a similar frequency of variants in the WES and healthy cohorts for LQTS1-3 (11.2% and 12.9%, respectively) and LQTS4-17 (27.1% and 38.4%, respectively). WES variants preferentially localized to amino acids altered in control individuals compared to cases. Based on amino acid-level analysis, WES-identified variants are indistinguishable from healthy background variation, whereas LQTS1 and 2 case-identified variants localized to clear pathologic "hotspots." No individuals who underwent clinical evaluation had clinical suspicion for LQTS.

Conclusion: The prevalence of incidentally identified LQTS-associated variants is ∼38% among WES tests. These variants most likely represent benign healthy background genetic variation rather than disease-associated mutations.

Keywords: Genetic testing; Genetics; Long QT syndrome; Mutation; Variant of undetermined significance; Whole exome sequencing.

Figure 1

A, Schematic of the study methodology. Synonymous variants, known polymorphisms, and variants interpreted as benign were excluded creating a proband-based WES cohort (light blue fill). Subset analysis of individuals seen a TCH with a variant in a LQTS1-3 gene (tan fill) was conducted. B, Pie chart of WES cohort demonstrating individuals with no LQTS-associated gene variants (negative, white fill) and individuals who are variant positive for LQTS1-3 (dark blue fill) and LQTS4-17 (light blue fill). C, Pie chart of WES variant-positive individuals depicting the number of variants hosted by a single proband.

Figure 2

Variant prevalence for LQTS, WES, and control cohorts. Bar graph of the frequency of LQTS-associated gene variants among individuals with LQTS (green fill), WES (light blue fill) and ostensible healthy control individuals (ExAC, light gray and GnomAD, dark gray fill). Error bars denote 95% CI. *, P<0.0001

Figure 3

WES and control cohort gene-specific variant prevalence. A, Bar graph of WES variants (blue fill) deemed “likely pathogenic” at time of genetic testing for each LQTS-associated gene. B, Variants deemed variants of undetermined significance (VUS). C, Rare variants among GnomAD control cohort (white fill). Missense (blue/white), intronic and untranslated regions (UTR, gray), and radical mutations (black) are noted. Error bars denoted 95% CI.

Figure 4

Gene-specific signal-to-noise frequencies. A, Bar graph of the frequency of LQTS/GnomAD-associated gene variants in the LQTS case cohort (green fill). B, Bar graph of the frequency of WES cohort/GnomAD-associated gene variants (blue fill). Yellow background denotes signal-to-noise threshold of 1. * Indicates a phenotype-positive genotype-negative proband sample. Error bars denote 95% CI.

Figure 5

Variant position overlap between LQTS, WES, and control cohorts. Venn diagram of the co-localization of variants to residues shared between the LQTS (green fill), control (white fill), and WES (blue fill) cohorts. The numbers within the circles demonstrate number of unique amino acid residues that host variants/mutations within each cohort. Numbers within overlapping portions of the circles reflect the number of residues with variants shared between respective cohorts. The proportion of shared variants, out of the cohort total, is noted within each box.

Figure 6

Amino-acid level signal-to-noise variant frequency for the three major gene products associated with LQTS. Frequency of LQTS (green) and WES (blue) cohort variant frequency versus amino acid position for Kv7.1 (A), Kv11.1 (B), and Nav1.5 (C). AKAP9, AKAP9-protein binding; CD, cytoplasmic; cNBD, C-terminal nucleotide binding; FGF, FGF13 binding; G, inactivation gate; KCNE1, KCNE1-protein binding; LC, L-type Ca²⁺ channel binding; PAS, Per-Arnt-Sim; SF, selectivity filter; S1–S6, transmembrane; SSTK, SSTK-interacting protein; TM, transmembrane; TSSK6-activating co-chaperone protein binding domains.

Figure 7

Pre- and post-test clinical suspicion for LQTS among TCH WES referrals. A, Bar graph of the frequency of indication by organ systems for WES referral. Individuals referred for cardiovascular (CV, blue fill). B, Schematic of the TCH cohort evaluation. Individuals seen at TCH with an ECG and no structural heart disease were included. C, Venn diagram of associated symptoms among individuals with QTc>460 (borderline) on at least 1 ECG.