Genomic Triangulation and Coverage Analysis in Whole-Exome Sequencing–Based Molecular AutopsiesClinical Perspective
Background—WEMA (Whole-Exome Molecular Autopsy) and surveillance of cardiac channelopathy and cardiomyopathy genes represents the latest molecular autopsy for sudden death in the young (SDY). To date, the majority of WEMA has been performed on the SDY case only.
Methods and Results—We performed whole-exome sequencing and nucleotide-level coverage analysis on 28 SDY cases (18.4±7.8 years) and their parents to determine the inheritance patterns of ultrarare, nonsynonymous variants in 99 sudden death–susceptibility genes. Nonsynonymous variants were adjudicated using the American College of Medical Genetics guidelines. Overall, 17 sudden death–susceptibility gene variants were identified in 12 of 28 (43%) SDY cases. On the basis of the American College of Medical Genetics guidelines, 6 of 28 (21%) cases had a pathogenic or likely pathogenic nonsynonymous variant with 3 (50%) being de novo. Two nonsynonymous variants would not have been elevated to likely pathogenic status without knowing their de novo status. Whole-exome sequencing reached a read depth of 10× across 90% of nucleotides within sudden death–susceptibility genes in 100% of parental exomes from fresh blood draw, compared with only 82% of autopsy-sourced SDY exomes.
Conclusions—An SDY-parent, trio-based WEMA may be an effective way of elucidating a monogenic cause of death and bringing clarity to otherwise ambiguous variants. If other studies confirm this relatively high rate of SDY cases stemming from de novo mutations, then the WEMA should become even more cost-effective given that the decedent’s first-degree relatives should only need minimal cardiological evaluation. In addition, autopsy-sourced DNA demonstrated strikingly lower whole-exome sequencing coverage than DNA from fresh blood draw.
Sudden cardiac death (SCD) among young, seemingly healthy individuals, is a tragic event for both families and their communities. Although the majority of SCD occurs in the elderly,1 annually, thousands of otherwise healthy individuals between the ages of 1 and 35 years die at an incidence of 1.3 per 100 000 persons.2 After a conventional autopsy, 40% of these cases remain inconclusive and are termed autopsy-negative sudden death in the young (SDY) or sudden unexplained death in the young,2 leaving families without answers surrounding the devastating loss of their young person.
See Editorial by Bartels et al
Autopsy-negative SDY may stem from cardiac channelopathies, such as long-QT syndrome, catecholaminergic polymorphic ventricular tachycardia, and Brugada syndrome, in which disturbed ion channel function can lead to lethal arrhythmias in the context of a structurally normal heart. In addition, cardiomyopathies, such as hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy, and arrhythmogenic cardiomyopathy, are known to cause SCD.3 Although the cardiomyopathies are typically associated with overt anatomic abnormalities (autopsy-positive SDY), some individuals manifest subtle features below a diagnosable level.4 Collectively, these potentially lethal cardiac diseases may underlie a substantial portion of both autopsy-negative SDY and autopsy-positive SDY cases, and there are currently 99 strong evidence or limited evidence SCD-susceptibility genes that have been identified.
Postmortem genetic testing (aka the molecular autopsy) of SDY victims for putative pathogenic variants in cardiac channelopathy- and cardiomyopathy-associated genes may reveal a plausible explanation for the death and is emerging as the new standard of care in decedent evaluation.5,6 Besides providing closure and clarity on the cause and manner of death in an SDY case, identifying a potential sudden death–causing variant may be crucial for surviving family members, as it may identify potentially at-risk relatives and may enable clinical interventions to prevent a subsequent tragedy. Whole-exome sequencing (WES) is an efficient and cost-effective approach to examine all 99 cardiac channelopathy- and cardiomyopathy-associated genes and when performed on a decedent, it is referred to as a WEMA (Whole-Exome Molecular Autopsy).7
WEMAs may be an effective approach for identifying potential sudden death–causing variants; however, to date, such molecular autopsies have been limited to the SDY case only.2,7–12 Performing WES on the SDY victims’ parents in addition to the victim (ie, parental genomic triangulation) reveals the inheritance pattern of identified variants that may be helpful in variant interpretation. In addition, previous WEMA studies have not comprehensively analyzed the quality of WES coverage within the cardiac channelopathy- and cardiomyopathy-associated genes when using postmortem sources of DNA. It is currently unknown whether using autopsy specimens (ie, frozen tissue, blood spot card, and autopsy blood) as a DNA source diminishes WES coverage quality.
Here, we sought to determine the use of an SDY-parent, trio-based WEMA for the identification of the genetic basis and inheritance pattern for sudden death in 28 young and seemingly healthy individuals whose sudden death was their sentinel event. We also performed a nucleotide-level coverage analysis to determine whether WES sufficiently captures variants within the sudden death–susceptibility genes and compared the coverage quality between DNA isolated from fresh blood (the parents) and autopsy material (the decedent).
Medical Examiner/Coroner-Referred SDY Cases
This WEMA study comprised 28 consecutive sudden death victims (64% male; 18.4±7.8 years; 96% white; 13 autopsy positive, 12 autopsy equivocal/negative, and 3 autopsy unavailable) who were referred to Mayo Clinic Windland Smith Rice Sudden Death Genomics Laboratory by medical examiners/coroners for postmortem genetic testing after a negative or inconclusive autopsy and their parents between August 2010 and September 2015 (Table 1).
Genomic DNA was isolated from autopsy whole blood, frozen tissue, or blood spot card for 15 of 28 (54%), 5 of 28 (18%), and 5 of 28 (18%) SDY victims, respectively, using the Gentra Puregene Blood Kit (Qiagen, MD) following the manufacturer’s protocol. Because the DNA was isolated elsewhere, the DNA source was unknown for 3 of 28 (11%) SDY victims. In addition, fresh peripheral blood leukocytes were obtained from each of the SDY victims’ parents following written informed consent in accordance with this Mayo Clinic Institutional Review Board approved study. DNA collection from both parents for all trios was available.
Next-Generation, Whole-Exome DNA Sequencing
Genomic DNA samples were submitted to Mayo Clinic Advanced Genomics Technology Center and Bioinformatics Core facility for WES. Paired-end libraries were prepared following Agilent protocol using the Bravo liquid handler. A Covaris E210 sonicator was used to fragment 1 to 3 µg of genomic DNA to 150 to 200 bp, ends were repaired, and an A base was added to the 3′ ends. Agilent paired-end Index DNA adaptors with a single T base overhang at the 3′ end were ligated, and the resulting constructs were purified using Agencourt AMPure SPRI beads. Adapter-modified DNA fragments were enriched by 4 cycles of polymerase chain reaction using SureSelect forward and Agilent SureSelect ILM Pre-Capture Indexing reverse primers. Agilent Bioanalyzer DNA 1000 chip determined the concentration and size distribution of the libraries.
Exon capture was performed using the Agilent SureSelect Human All Exon V5+Untranslated Region Target Enrichment System. Whole-exon biotinylated RNA capture baits were incubated with 750 ng of the prepped library for 24 hours at 65°C. Captured DNA:RNA hybrids were recovered with Dynal Dynabeads MyOne Streptavidin T1, and DNA was eluted from the beads and purified using Agencourt Ampure XP beads. Agilent Sure Select postcapture indexing forward and index polymerase chain reaction reverse primers were used to amplify purified capture products.
Exome libraries were loaded onto paired-end flow cells at equimolar concentrations of 7 to 8 pmol/L to generate cluster densities of 600 000 to 800 000/mm2 using the Illumina cBot and HiSeq Paired-end cluster kit version 3, following Illumina protocol. Each lane of a HiSeq flow cell produced 21 to 39 Gbases of sequence, and the level of sample pooling was controlled by the size of the capture region and the desired depth of coverage. The flow cells were sequenced as 101× 2 paired-end reads on an Illumina HiSeq 2000 using TruSeq SBS sequencing kit version 3 and HiSeq data collect version 220.127.116.11 software, and base-calling was performed using Illumina RTA version 18.104.22.168.
The Illumina paired-end reads were aligned to the hg19 reference genome using BWA-MEM followed by the sorting and marking of duplicate reads with Picard (http://picard.sourceforge.net). Genome Analysis Toolkit was used for local realignment of insertions/deletions and base quality score recalibration. Single-nucleotide variants and insertions/deletions were called across all of the samples simultaneously using Genome Analysis Toolkit HaplotypeCaller and GenotypeGVCFs with variant quality score recalibration.13 The expected relatedness of each parent pair (unrelated) and each parent–offspring (related) was confirmed using an identity by descent probability analysis that was performed using PREST-plus.14
Variant Filtering and Pathogenicity Assessment
After WES, variants were filtered using Qiagen Ingenuity Variant Analysis software using sporadic de novo, dominant, and recessive (ie, compound heterozygous or homozygous) inheritance models. Variants were included only if they met the following filtering parameters: (1) had a high-quality score (read depth >10 reads, call quality >20, genotype quality >20, and present in genes outside the top 1% of exonically variable genes and top 5% of exonically variable 100 base windows), (2) were nonsynonymous (ie, missense, nonsense, frameshift insertion/deletion [INDEL], in-frame INDEL, or splice error), and (3) met our rarity thresholds (minor allele frequency ≤0.00005 for de novo and dominant heterozygous models; ≤0.005 for recessive homozygous and compound heterozygous models in Exome Aggregation Consortium [n=60 706],15 1000 Genome Project [n=1094],16 and the National Heart, Lung and Blood Institute Grand Opportunity Exome Sequencing Project [n=6503]17 databases). Variants meeting the above criteria underwent a further gene-specific surveillance for all known cardiac channelopathy-susceptibility, cardiomyopathy-susceptibility, and sudden unexplained death in epilepsy–susceptibility genes (n=99, Table I in the Data Supplement).
All identified variants and inheritance models were confirmed using Sanger sequencing with standard polymerase chain reaction and DNA dye terminator cycle sequencing protocols and an ABI Prism 377 automated sequencer (Applied Biosystems Inc, Foster City, CA). DNA sequence chromatograms were analyzed using Chromas version 1.45 and Gene Codes Corporation Sequencher version 5.4.1 sequence analysis software.
Subsequently, all identified variants were analyzed using the following mutation prediction in silico tools: polymorphism phenotyping,18 Protein Variation Effect Analyzer,19 Sorting Intolerant From Tolerant,20 Mutation Assessor, Functional Analysis through Hidden Markov models,21 and Combined Annotation-Dependent Depletion.22
The pathogenicity of all identified variants within the sudden death–associated genes was determined using the American College of Medical Genetics and Genomics (ACMG) standards and guidelines.23 Briefly, the guidelines adjudicate variants as benign, likely benign, variant of uncertain significance (VUS), likely pathogenic, or pathogenic based on a point system. Variants accumulate points by meeting specific criteria, such as previously published data, inheritance pattern, prevalence in large public databases, and in silico mutation prediction tools. The ACMG recommends that variants classified as either likely pathogenic or pathogenic be reported to family members and should therefore be considered clinically actionable.
Depth of Coverage Evaluation
Binary Alignment Map files from WES were used to perform a gene-specific survey of 99 channelopathy-susceptibility (long-QT syndrome, catecholaminergic polymorphic ventricular tachycardia, and Brugada syndrome), cardiomyopathy-susceptibility (HCM, dilated cardiomyopathy, and arrhythmogenic cardiomyopathy), and sudden unexplained death in epilepsy–susceptibility genes (Table I in the Data Supplement). These include the 6 major contributor genes KCNQ1, KCNH2, SCN5A, RYR2, MYH7, and MYBPC3. Gene and exon coordinates for the 99 genes were obtained from Ensembl release 75 downloaded from the Ensembl FTP resource (http://ftp.ensembl.org/pub/release-75/gtf/homo_sapiens/). Genome Analysis Toolkit DepthOfCoverage walker was used to calculate the coverage of the gene regions from the aligned Binary Alignment Map files. Read depths of 10×, 20×, and 30× (at least 10, 20, and 30 sequence reads aligned at a nucleotide position, respectively) were used as thresholds. Nucleotides with <10× coverage were considered to have poor coverage. Union of the gene’s Ensemble transcripts was used to define genes with good coverage as those having >90% of the region mapped with a read depth of at least 10×. Poorly covered genes and exons were listed and further evaluated. The numbers of nucleotides with <10 reads mapped were evaluated separately for all 99 genes. Statistical comparison of average coverages of parents and SDY cases at 10×, 20×, and 30× was performed using 2-tailed t tests assuming unequal variances (df=27; t= 2.4, 2.8, 3.0 for 10×, 20×, 30×, respectively). Statistical comparison of proportions of parents and SDY with >90% of nucleotides covered was performed using an χ2 test of independence (df=1; χ2=10.6, 13.4, 18.5 for 10×, 20×, 30×, respectively).
This WEMA study comprised 28 consecutive sudden death victims who were referred between August 2010 and September 2015 to Mayo Clinic Windland Smith Rice Sudden Death Genomics Laboratory by medical examiners/coroners for postmortem genetic testing after a negative or inconclusive autopsy. Following written informed consent in accordance with this Mayo Clinic Institutional Review Board approved study, WES was performed on genomic DNA isolated from autopsy whole blood, frozen tissue, blood spot card, or unknown source for 15 of 28 (54%), 5 of 28 (18%), 5 of 28 (18%), and 3 of 28 (11%) SDY victims. DNA was isolated from fresh peripheral blood leukocytes for each SDY victim’s parents.
Detailed cohort demographics and characteristics are displayed in Table 1. The cohort consisted of 28 SDY victims (64% male; 18.4±7.8 years; 96% white and 4% Indian). There were 18 males (17.5±9.7 years, range 1.3–38 years) and 10 females (18.8±4.7 years, range 15–25 years). The 28 SDY victims died during the following events: sleep (18; 64%), unwitnessed (5; 18%), exertion (running, swimming, wakeboarding, and emotion [4; 14%]), and nonspecific (1; 4%). Conventional cardiac evaluation at autopsy did not identify any gross or microscopic structural abnormalities in 7 (25%) cases (ie, autopsy-negative SDY). Five (18%) SDY victims had equivocal structural abnormalities including myocardial bridging of the left anterior descending coronary artery, scattered hypereosinophilic cardiomyocytes, dispersion of His bundle, and fatty infiltration of atrioventricular node. However, these findings were not deemed conclusive. Thirteen SDY cases (46%) were autopsy-positive SDY cases with gross/microscopic evidence of pathological cardiac hypertrophy, fibrosis, disarray, and dilatation (ie, a suspected cardiomyopathy). Autopsy reports were unobtainable for 3 (11%) SDY victims.
Yield and Inheritance Pattern of Ultrarare Nonsynonymous Variants in Sudden Death–Susceptibility Genes
On average, each SDY victim hosted 47.2±22.5 (range 6–100) dominantly inherited (maternal or paternal) and 3.1±4.0 (range 0–17) de novo ultrarare (minor allele frequency <0.00005) nonsynonymous variants (NSVs) across their entire exome. After the 99 gene-specific surveillance (Table I in the Data Supplement), 17 sudden death–susceptibility gene NSVs were identified in 12 of 28 (43%) SDY cases (Figure 1; Table 1). Three SDY victims hosted multiple variants. Of the variant-positive SDY cases, genomic triangulation revealed that 3 of 12 (25%) hosted de novo NSVs, whereas 9 of 12 (75%) hosted dominantly inherited NSVs. In SDY victims with signs of cardiac hypertrophy/dilation, 7 of 13 (54%) hosted a NSV; compared with the 4 of 12 (33%) found in victims with structurally normal hearts or equivocal findings at autopsy.
The pathogenicity of all identified variants within the sudden death–associated genes was determined using the ACMG standards and guidelines.23 After careful interpretation and adjudication of NSVs based on the strict ACMG guidelines (Table 2), 6/28 (21%) cases had a pathogenic (p.L39X-PLN, p.R454W-DES, p.Y392X-ABCC9, p.F111fs-BAG3) or likely pathogenic (p.T857P-MYH7, p.L113P-RYR2) NSV. Of the SDY victims with autopsy-positive SDY, 4 of 13 (31%) had either a pathogenic or likely pathogenic NSV compared with 1 of 12 (8%) of the autopsy-negative/equivocal SDY cases. Of the 6 NSVs adjudicated as pathogenic or likely pathogenic, 3 (50%) were de novo (Figure 1; Table 1). Importantly, two of these NSVs (p.T857P-MYH7, p.L113P-RYR2) would not have been elevated from the classification level of VUS to likely pathogenic without knowing their de novo status.
A maternally derived p.L39X-PLN pathogenic NSV was identified in a 21-year-old white male with a 500-g heart and dilation of all 4 chambers at autopsy (Table 1, case 20). The p.L39X-PLN variant has been associated previously with HCM and dilated cardiomyopathy.24
A de novo p.R454W-DES pathogenic NSV was identified in a 13-year-old white male with a 430-g heart, a 1.7-cm-thick left ventricle wall, left ventricle myofibril disarray, and fibrosis in the left atrium and papillary muscles (Table 1, case 5). The p.R454W-DES was described previously as a de novo pathogenic mutation in a 15-year-old male who developed exercise intolerance and HCM leading to cardiac transplantation at age 25 years. Both in vivo and in vitro studies of p.R454W-DES demonstrated a dramatic effect on filament formation.25,26
Another pathogenic NSV, a paternally derived p.Y392X-ABCC9, was found in a 16-year-old white male with a 500-g heart, a 1.7-cm left ventricle wall, and myocyte hypertrophy (Table 1, case 8). All aforementioned SDY victims died during sleep. A pathogenic NSV, maternally derived p.F111fs-BAG3, was identified in a 25-year-old white female who drowned while swimming (Table 1, case 24).
A likely pathogenic, de novo p.L113P-RYR2 NSV was hosted by a 21-year-old Indian male who died during emotion. Nonspecific dispersion of the His bundle and fatty infiltration of the atrioventricular node was noted during his autopsy (Table 1, case 19).
Finally, a de novo, likely pathogenic NSV, p.T857P-MYH7, was identified in a 16-year-old white male who died while running. At autopsy, he was diagnosed with HCM (480-g heart, septal wall thickness of 4.0 cm, and myofiber disarray and fibrosis in the interventricular septum; Table 1, case 9).
WES Coverage Analysis of Sudden Death–Susceptibility Genes
Binary Alignment Map files from WES were used to perform a gene-specific survey of 99 channelopathy-susceptibility (long-QT syndrome, catecholaminergic polymorphic ventricular tachycardia, and Brugada syndrome), cardiomyopathy-susceptibility (HCM, dilated cardiomyopathy, and arrhythmogenic cardiomyopathy), and sudden unexplained death in epilepsy–susceptibility genes (Table I in the Data Supplement). The nucleotide-level coverage analysis revealed that 84 of the 99 (85%) sudden death–susceptibility genes had good overall coverage in 28 SDY victims and 56 parental DNA exomes, with >90% of nucleotides having a read depth of at least 10× across all exons. However, 15 of 99 (15%) genes contained at least 1 exon with poor coverage (<90% of nucleotides covered at 10×; Table 3). Eleven of these (11% overall) genes contained exons with extreme poor coverage (<5% of nucleotides covered at 10×; Table 3). At read depths of 20× and 30×, the number of genes with poorly covered exons increased to 14 and 20, respectively, including the major cardiac channelopathy and cardiomyopathy genes: KCNQ1, RYR2, and MYH7 (data not shown).
In addition, the coverage analysis demonstrated significant differences in the percentage of bases covered between the exomes of SDY victims and their parents. All parental DNA was obtained via a fresh blood draw, whereas the SDY DNA was extracted from a variety of postmortem sources: whole blood in 15, frozen tissue in 5, blood spot card in 5, and unknown in 3. At read depths of 10×, 20×, and 30×, the parents’ exome sequences derived from fresh blood-draw DNA had average coverages of 98%, 96%, and 93%, whereas SDY victims’ exome sequences derived from autopsy material DNA had average coverages of 92%, 87%, and 83% with P values of 0.02, 0.009, and 0.006, respectively.
Given the relative sample sizes, none of the postmortem sources of DNA outperformed the other statistically in terms of yielding better DNA coverage. DNA derived from blood cards (n=5 cases) had average coverages of 97%, 93%, and 89% at read depths of 10×, 20×, and 30×, respectively. Frozen tissue-derived DNA (n=5 cases) had average coverages of 93%, 90%, and 86%. In autopsy whole blood-derived DNA (n=15 cases), 88%, 83%, and 78% of nucleotides within sudden death–susceptibility genes were covered at read depths of 10×, 20×, and 30×, respectively.
The proportion of parents’ exome sequences with read depths of at least 10×, 20×, and 30× across 90% of nucleotides (good coverage) within the 99 sudden death–susceptibility genes was 56 of 56 (100%), 54 of 56 (96%), and 49 of 56 (88%), respectively. However, the proportion among SDY victims’ exome sequences was only 23 of 28 (82%, P=0.001), 19 of 28 (68%, P=0.0003), 15 of 28 (54%, P=0.0001; Figure 2), demonstrating substantially lower quality of exome coverage from DNA isolated from autopsy material compared with a fresh blood draw. The proportions of autopsy whole blood-derived DNA samples (n=15 cases) with at least 90% of nucleotides covered were 11 of 15 (73%), 9 of 15 (60%), and 7 of 15 (47%) at read depths of 10×, 20×, and 30×, respectively. For DNA derived from frozen tissue (n=5 cases), the proportions was 1 of 5 (20%) for all read depths, and for blood card–derived DNA (n=5 cases), the proportions were 5 of 5 (100%), 4 of 5 (80%), and 2 of 5 (40%) for read depths of 10×, 20×, and 30×, respectively.
Over the past decade, various standardizations for postmortem evaluation of SDY have been proposed. In 2008, the Association for European Cardiovascular Pathology recommended the use of postmortem genetic analysis for both structural (autopsy-positive SDY) and nonstructural heart disease (autopsy-negative SDY).27 In addition, the members of the TRAGADY (Trans-Tasman Response Against Sudden Death in the Young) put forward guidelines to ensure standardization of autopsy practices for SDY and the preservation of appropriate DNA-friendly material for all cases of SCD to be used for genetic testing.28 In 2011, the Heart Rhythm Society and the European Heart Rhythm Association recommended genetic testing in SDY cases, especially where clinically relevant presentations existed among the decedent or family members.6 However, because of the cost and time-consuming nature of current genetic testing efforts, this type of postmortem investigation has been very difficult for medical examiners and coroners to provide.
WES is a cost-effective method for facilitating the comprehensive analysis of a patient’s entire library of protein-coding genes. The use of WES as an approach for postmortem molecular autopsy among SDY cases has been recognized increasingly.2,7–12 Here, using WES and a gene-specific analysis involving 99 cardiac channelopathy-associated, cardiomyopathy-associated, and sudden unexplained death in epilepsy–associated genes, we identified an ultrarare (minor allele frequency <0.00005; 1 in 20 000 alleles), amino acid altering, sudden death–susceptibility gene–associated variant in 43% of our SDY victims. However, the challenge of elucidating the underlying genetic cause of SDY lies not in the identification of ultrarare NSVs, but in the interpretation of their potential pathogenicity.
Although there is increasing evidence to support the use of WES, standardization for characterizing putative pathogenic mutations is of paramount importance to enable proper counseling of surviving family members.29,30 In addition, standardized adjudication is crucial for accurate publishing and scientific progress. Besides the various underlying phenotypes beneath SDY, much of the discrepancy in reported yields of variant-positive cases among molecular autopsy studies stem from differences in the variant adjudication process (ie, mutation calling). In 2015, the ACMG and the Association for Molecular Pathology provided standards and guidelines for the interpretation of sequence variants.23 These criteria for the determination of pathogenicity should be applied in delineating the pathogenicity of variants identified by WEMA among SDY cases.
Many of the NSVs within sudden death–associated genes identified in previously reported molecular autopsy series do not meet the strict criteria put forth in the ACMG guidelines for classification as pathogenic or likely pathogenic.2,7–12 In other words, many of the SDY-associated variants may have been classified prematurely and erroneously as a disease-causative mutation. In our current study, while we can claim that 43% of SDY victims hosted at least 1 ultrarare NSV, to suggest that we have explained 43% of these SDY cases with a monogenetic lesion would be wrong and misleading. Instead, with strict application of the ACMG guidelines for variant adjudication, 21% of our cases hosted pathogenic or likely pathogenic variants that would be deemed clinically actionable. Some identified variants that do not meet the ACMG criteria of pathogenicity may in fact be disease causing. However, given the potential implications if applied prematurely for cascade testing of the living relatives, they should remain classified as VUS until reaching a sufficient level of evidence for pathogenicity through the use of functional validation studies and other criteria set forth in the guidelines.
The ACMG guidelines provide a set of criteria which enable sequence variants to be stratified into varying levels of pathogenicity based on a weighted point system (Table 2).23 Identified variants are assessed for very strong (ie, a null variant in a gene where loss of function is a known mechanism of disease), strong (ie, the same amino acid change as a previously established pathogenic variant, confirmed as de novo when there is no family history, or is associated with well-established functional studies demonstrating a deleterious effect), moderate (ie, absence from control populations in large public databases like Exome Aggregation Consortium, located in a well-established functional domain), or supporting (ie, multiple lines of computational evidence predicting a deleterious effect) evidence for pathogenicity. Points earned in each category are combined in a variety of ways to reach a final variant classification level of pathogenic, likely pathogenic, or VUS. As the incorporation of postmortem genetic testing as part of a truly comprehensive autopsy continues to rise, it is crucial that these ACMG guidelines be implemented to minimize erroneous variant adjudication. Perhaps the only thing worse than telling a family that we do not know why their loved one died is to tell them that we have found the genetic answer when we have not.
The use of genomic triangulation, or performing WES on the SDY victims as well as their parents, offers insights into the potential pathogenicity of an identified variant by revealing immediately its inheritance model. This is essential for comprehensive variant interpretation and adjudication based on the ACMG guidelines and informs the necessity of subsequent clinical evaluations for surviving family members. According to the ACMG guidelines, a de novo variant in an affected individual with no family history of disease provides strong evidence of pathogenicity. In our study, without using the SDY-parent, trio-based WEMA approach, 4 of 28 (14%) SDY victims hosted pathogenic or likely pathogenic NSVs, whereas 8 of 28 (29%) victims hosted ultrarare NSVs within cardiac channelopathy or cardiomyopathy genes that were classified as VUS. However, genomic triangulation revealed that 2 of the 8 SDY victims with an apparent VUS could have their variant promoted to likely pathogenic given its de novo status.
In addition, the identification of sudden death–causing NSVs using genomic triangulation enables instant risk stratification for surviving family members in some cases. In our study, 3 of 6 (50%) pathogenic NSVs were de novo, revealing that none of the surviving family members are at risk of SCD mediated by the same mechanism as the decedent (of course with the caveat of gonadal mosaicism being considered). If other studies confirm this relatively high rate of de novo mutation-mediated SDY, then the decedent’s sibling would require minimal cardiological evaluation and variant-specific testing just in case one of the parents possessed the variant as a gonadal mosaic, and WEMA should become even more cost-effective.
Recently, guidelines for autopsy investigations of SDY cases stipulate procurement and retention of tissue suitable for DNA extraction as a class I recommendation and advise that postmortem genetic testing be considered as the new standard of care in the decedent’s evaluation.5,6,28 However, until this study, it has been unknown whether using autopsy specimens (ie, frozen tissue, blood spot card, or autopsy blood) as a DNA source diminishes WES coverage quality compared with freshly procured peripheral blood lymphocytes from a living host. Sufficient capture quality at the gene, exon, and nucleotide level within sudden death–susceptibility genes is necessary for WES to be considered an effective approach for postmortem genetic testing of SDY.
Although WEMA successfully elucidated the underlying genetic cause of death in multiple SDY cases, our detailed coverage analysis of every nucleotide residing within every exon of these 99 analyzed genes exposed deficiencies in the WES of DNA sourced from autopsy material. Among parental DNA obtained from fresh blood draws, WES sufficiently captured nearly all protein-coding regions of the sudden death–susceptibility genes, with 90% of nucleotides reaching a read depth of 10× in 100% of parents. However, this threshold was met in only 82% of SDY victims, demonstrating the strikingly lower coverage of WES with autopsy-sourced DNA. Clearly, this decreased coverage could result in potentially lethal variants escaping detection. Our preliminary comparison of autopsy materials suggests that blood cards may be a better DNA source for WES than whole blood (which has been used exclusively in 2 recent studies2,8) or frozen tissue; however, this is based on small sample sizes. WES of blood cards covered 90% of nucleotides at 10× in 100% of SDY victims, whereas whole blood and frozen tissue reached this threshold of coverage in only 73% and 20% of cases, respectively. However, the autopsy materials were obtained and stored in a variety of medical examiners/coroners offices, so these differences may not be because of the autopsy material alone. Methods of collection and preservation may also contribute to the ability of WES to effectively capture the sudden death–susceptibility genes in their entirety. Further study with larger sample sizes are needed to determine the preferred autopsy material source, method of collection, and storage of DNA from decedents for WES.
Parental genomic triangulation in WEMA of SDY revealed that ≈20% of these SDY cases hosted pathogenic variants with half being de novo. An SDY-parent, trio-based WEMA may be an effective strategy of elucidating a monogenic cause of sudden death and bring clarity to otherwise ambiguous variants. If other studies with larger sample sizes confirm this relatively high rate of SDY stemming from de novo mutations, then the WEMA should become even more cost-effective given that the decedent’s family members would require minimal cardiological evaluation when the pathogenic cause of death is ascribed to a de novo mutation
We gratefully acknowledge both the medical examiners and coroners for referring the sudden death victims to our program in an effort to find an explanation for their sudden death and the parents of the decedents of the victims for participating with us to search for answers to their family’s tragedy.
Sources of Funding
This work was supported by the Mayo Clinic Windland Smith Rice Comprehensive Sudden Cardiac Death Program.
Dr Ackerman is a consultant for Audentes Therapeutics, Boston Scientific, Gilead Sciences, Invitae, Medtronic, MyoKardia, and St. Jude Medical. D.J. Tester, Dr Ackerman, and Mayo Clinic have received royalties from Transgenomic for their FAMILION-LQTS and FAMILION-CPVT genetic tests. Dr Ackerman and Mayo Clinic have an equity/royalty-based licensing agreement with AliveCor. However, none of these entities were involved in this study in any way. The other authors report no conflicts.
The Data Supplement is available at http://circgenetics.ahajournals.org/lookup/suppl/doi:10.1161/CIRCGENETICS.117.001828/-/DC1.
- Received May 17, 2017.
- Accepted July 21, 2017.
- © 2017 American Heart Association, Inc.
- Priori SG,
- Wilde AA,
- Horie M,
- Cho Y,
- Behr ER,
- Berul C,
- et al
- Ackerman MJ,
- Priori SG,
- Willems S,
- Berul C,
- Brugada R,
- Calkins H,
- et al
- Anderson JH,
- Tester DJ,
- Will ML,
- Ackerman MJ
- Christiansen SL,
- Hertz CL,
- Ferrero-Miliani L,
- Dahl M,
- Weeke PE,
- Santori M,
- Blanco-Verea A,
- Gil R,
- Cortis J,
- Becker K,
- Schneider PM,
- et al
- Sun L,
- Dimitromanolakis A
- Auer PL,
- Johnsen JM,
- Johnson AD,
- Logsdon BA,
- Lange LA,
- Nalls MA,
- et al
- Richards S,
- Aziz N,
- Bale S,
- Bick D,
- Das S,
- Gastier-Foster J,
- et al
- Brodehl A,
- Hedde PN,
- Dieding M,
- Fatima A,
- Walhorn V,
- Gayda S,
- et al
The WEMA (Whole-Exome Molecular Autopsy) with surveillance of the cardiac channelopathy and cardiomyopathy genes represents the latest molecular autopsy for sudden death in the young (SDY). To date, the majority of WEMA has been performed on the decedent only. We performed whole-exome sequencing and nucleotide-level coverage analysis on 28 SDY cases and their parents to determine the inheritance patterns of ultrarare, nonsynonymous variants in 99 sudden death–susceptibility genes. Variants were adjudicated using the American College of Medical Genetics guidelines. Overall, rare sudden death–susceptibility gene variants were identified in 43% of SDY cases. On the basis of the American College of Medical Genetics guidelines, 21% of cases had a pathogenic or likely pathogenic variant with half being de novo variants. Two of the variants would not have been elevated to likely pathogenic status without knowing their de novo status. An SDY-parent, trio-based WEMA may be an effective way of elucidating a monogenic cause of death and bringing clarity to otherwise ambiguous variants. If other studies confirm this relatively high rate of SDY cases stemming from de novo mutations, then the WEMA should become even more cost-effective given that the decedent’s first-degree relatives should only need minimal cardiological evaluation. In addition, autopsy-sourced DNA demonstrated strikingly lower whole-exome sequencing coverage than DNA from fresh blood draw. Further study with larger sample sizes are needed to determine the preferred autopsy material source, method of collection, and storage of DNA from decedents for whole-exome sequencing.