Article Text

Download PDFPDF

Short report
Rapid detection of genetic variants in hypertrophic cardiomyopathy by custom DNA resequencing array in clinical practice
  1. Siv Fokstuen1,
  2. Analia Munoz2,
  3. Paola Melacini3,
  4. Sabino Iliceto3,
  5. Andreas Perrot4,
  6. Cemil Özcelik4,
  7. Xavier Jeanrenaud5,
  8. Claudine Rieubland6,
  9. Martin Farr7,
  10. Lothar Faber7,
  11. Ulrich Sigwart8,
  12. François Mach8,
  13. René Lerch8,
  14. Stylianos E Antonarakis1,2,
  15. Jean-Louis Blouin1
  1. 1Genetic Medicine, University Hospitals of Geneva, Geneva, Switzerland
  2. 2Department of Genetic Medicine and Development, University of Geneva School of Medicine, Geneva, Switzerland
  3. 3Department of Cardiac, Thoracic and Vascular Sciences, University of Padua, Padua, Italy
  4. 4Cardiology and Experimental & Clinical Research Center, Charité-Universitaetsmedizin Berlin, Germany
  5. 5Cardiology, University Hospitals of Lausanne, Lausanne, Switzerland
  6. 6Division of Human Genetics, Department of Paediatrics, University of Bern, Bern, Switzerland
  7. 7Kardiologische Klinik, Herz- und Diabeteszentrum NRW, Bad-Oeynhausen, Germany
  8. 8Cardiology, University Hospitals of Geneva, Geneva, Switzerland
  1. Correspondence to Dr Siv Fokstuen, Genetic Medicine, Centre Médical Universitaire, 1 rue Michel-Servet, 1211 Geneva 4, Switzerland; siv.fokstuen{at}


Background Hypertrophic cardiomyopathy (HCM) is the most common inherited cardiac disease (1/500) and the most common cause of sudden cardiac death in young people. Pathogenic mutation detection of HCM is having a growing impact on the medical management of patients and their families. However, the remarkable genetic and allelic heterogeneity makes molecular analysis by conventional methods very time-consuming, expensive and difficult to realise in a routine diagnostic molecular laboratory.

Method and results The authors used their custom DNA resequencing array which interrogates all possible single-nucleotide variants on both strands of all exons (n=160), splice sites and 5′-untranslated region of 12 HCM genes (27 000 nucleotides). The results for 122 unrelated patients with HCM are presented. Thirty-three known or novel potentially pathogenic heterozygous single-nucleotide variants were identified in 38 patients (31%) in genes MYH7, MYBPC3, TNNT2, TNNI3, TPM1, MYL3 and ACTC1.

Conclusions Although next-generation sequencing will replace all large-scale sequencing platforms for inherited cardiac disorders in the near future, this HCM resequencing array is currently the most rapid, cost-effective and reasonably efficient technology for first-tier mutation screening of HCM in clinical practice. Because of its design, the array is also an appropriate tool for initial screening of other inherited forms of cardiomyopathy.

  • Hypertrophic cardiomyopathy
  • genetic testing
  • resequencing array
  • mutation
  • cardiovascular medicine
  • cardiomyopathy
  • diagnostics tests
  • genetic screening/counselling
  • molecular genetics
View Full Text

Statistics from


The wide genetic heterogeneity of inherited cardiac disorders and the laborious, time-consuming and expensive serial molecular methods make it very difficult to unravel rapidly the causative mutation in a routine diagnostic molecular laboratory. The tremendous potential of genetic testing in clinical practice has therefore so far often been limited by the lack of efficient screening methods. Hypertrophic cardiomyopathy (HCM) represents the most common inherited cardiac disorder (prevalence of 1:500), and is also thought to be the primary cause of sudden cardiac death (SCD) in young adults and competitive athletes.1 The disease is usually familial, with an autosomal dominant mode of inheritance and a very wide inter- and intra-familial clinical variability.2

Over the last two decades, several hundred pathogenic mutations in at least 21 different HCM susceptibility genes have been identified.3 The most common form of HCM is caused by alterations in genes of the cardiac sarcomere. More recently, mutations in additional, non-sarcomeric genes have been identified encoding various calcium-handling (phospholamban, junctophilin-2) and Z-disc proteins (CSRP3-encoded muscle LIM protein, TCAP-encoded telethonin, LDB3-encoded LIM domain binding 3, ACTN2-encoded α-actinin 2, VCL-encoded vinculin/metavinculin, MYOZ2-encoded myozenin-2).3 The majority of HCM mutations (87%) are single-nucleotide substitutions. The remaining 13% are small insertions or deletions, mainly in MYBPC3, or rare large deletions. In 3–5% of the families, double mutations have been found.4 These cases usually have a more severe clinical presentation. Interestingly, in most cases of digenic inheritance, one of the mutations usually involves MYBPC3.5

In order to overcome the laborious and expensive conventional mutation screening approach for such a heterogeneous condition, we have developed a 27 kbp custom DNA resequencing array for 12 HCM genes.6 The resequencing array is very efficient for the detection of nucleotide substitutions, but limited with regard to quantitative analysis of deletions/insertions. After the technical validation of the array in 2008,6 we analysed 122 unrelated patients with HCM from six different centres. Our results highlight that the HCM resequencing array is currently the most rapid, cost-effective and reasonably efficient method for mutation screening of HCM in clinical practice.

Materials and methods


With approval from the local ethics committee, we analysed 122 unrelated patients with HCM (64 with a positive family history and 58 with unremarkable or unknown family history) from six different centres. All patients (57 from Geneva, eight from Lausanne, one from Berne, Switzerland, 35 from Padua, Italy, 13 from Berlin, eight from Bad Oeynhausen, Germany) were selected by cardiologists according to the European diagnostic criteria for familial HCM,2 and written informed patient consent was obtained. We included in this series of 122 patients the 38 patients previously reported on in 2008.6

For the assessment of unreported potentially pathogenic non-synonymous variants, we used DNA samples from healthy Caucasian people as controls (96 samples from the Charité-Universitätsmedizin Berlin, 150 samples from Padua). Cardiomyopathy was excluded by echocardiography in all of these controls, and their mean age was 70.8 years.

Resequencing array

DNA was isolated from lymphocytes using standard protocols. We analysed all 122 patients in Geneva by our previously described DNA resequencing array covering all coding exons (n=160), splice-site junctions, and 5′-untranslated region (UTR) of both strands of 12 HCM genes (MYH7, MYBPC3, TNNT2, TPM1, TNNI3, MYL3, MYL2, CSRP3, PLN, ACTC1, TNNC1, PRKAG2).6 The sequence was determined on both strands. Putative functional DNA sequence variants were all confirmed by Sanger sequencing. All variants found in gene MYH7 were first evaluated by BLAST analysis in order to avoid eventual misinterpretation with the paralogous MYH6 gene.

Previously unreported, non-synonymous confirmed sequence variants were further assessed for potential deleteriousness by screening a healthy control population, or more recently by consulting the 1000-Genomes browser (


HCM array sequence performance

After hybridisation of the 122 arrays, we had a mean nucleotide call rate of 98.4% using Affymetrix GDAS/GSEQ software pipeline (range 93.0–99.9%). We observed a constant improvement of the call rate as the number of experiments increased. We had ∼1.6% of unread sequence (no calls) per array, mainly observed in regions with two to six consecutive C bases, reflecting suboptimal hybridisation of CpC-rich areas. So far, no HCM mutations have been reported in such CpC-rich regions (HGMD, The following exons were particularly rich in such unread nucleotides: MYBPC3, exon 3; TPM1, exon 4; ACTC1, exon 1; PRKAG2, exon 3.

Variants identified by the HCM array

In total, we identified 34 heterozygous known or novel potentially pathogenic variants in 42/122 patients (34%) (table 1). Sixteen (two nonsense variants and 14 missense variants) were recorded in the HGMD database (table 1). Four variants were novel splice-site mutations in MYBPC3. Twelve missense variants and two 5′-UTR variants were novel changes. Eleven of the novel missense variants and the MYH7 5′-UTR change were either absent from at least 192 Caucasian control chromosomes or were not recorded in the 1000-Genomes database. Sequence alignment among different mammals and vertebrates revealed that they affect highly conserved residues. Furthermore, we identified five known non-synonymous single-nucleotide polymorphisms (SNPs) and 30 known synonymous SNPs (table 2). Four of the known non-synonymous SNPs were initially reported as disease-causing variants.20–23 Although it now seems clear that they are common variants in the general population, a modifying role in the pathogenesis of HCM cannot be ruled out. The role of the variant p.Ser175Gly in exon 33 of MYH7 (table 2) is less clear. As reported in dbSNP (rs2754155), its frequency is very low (MAF=0.006).

Table 1

Heterozygous known or novel potentially pathogenic variants

Table 2

Variants known as synonymous or non-synonymous single-nucleotide polymorphisms

According to these results, at least 33 of the 69 heterozygous variants can be considered as known pathogenic mutations or as novel very likely pathogenic variants. They were found in 38 different probands, which corresponds to a total mutation detection rate of 31% (38/122). Twenty-eight had a positive family history of SCD or HCM. Considering only the familial cases, the mutation detection rate was 45% (28/62). We did not find any patient with more than one mutation.

We observed considerable clinical variability with regard to age at diagnosis and course of the disease. However, our cohort is too small to establish meaningful genotype–phenotype correlations. One patient with early-onset HCM and a positive family history of SCD had a potentially pathogenic variant in 5′-UTR of MYH7 (c.1-2127A→T, NM_000257.2). To our knowledge, no pathogenic variant has so far been reported in the 5′-UTR of a sarcomeric gene. Our variant was recorded as a SNP (rs72686233) in March 2006 assembly and in dbSNP(130). However, this 5′-UTR variant was omitted from the dbSNP(131). According to the UCSC browser, it appears to be conserved down to Opossum, and it is predicted to be at a promoter histone mark (H3K4Me3) site in HepG2 cells.


Because of extensive genetic and allelic heterogeneity, molecular diagnosis of HCM by conventional serial methods is problematic. The usual experience of the patient and his/her family is that a few laboratories provide a small number of tests with a long, often unpredictable, turnaround time and very high costs. Systematic genotyping of nine sarcomeric genes revealed that more than 80% of the disease-causing mutations are found in MYH7, MYBPC3 and TNNT2.8 Thus, genetic testing in an index case usually starts with mutation analysis of these three most commonly affected genes eventually followed, depending on the facilities of the laboratories, by the analysis of other sarcomeric genes. To overcome the great genetic heterogeneity of HCM and to improve the diagnostic possibilities, we developed a 12-gene DNA custom resequencing array for HCM.6 In the same year, Waldmüller et al published the design of a resequencing array covering the three most commonly affected HCM genes, MYH7, MYBPC3 and TNNT2.24 Shortly after our publication, a few companies (Harvard Partners, Correlagen, PGxHealth and GeneDx) offered testing for up to 17 HCM genes by array technology, and a custom resequencing array containing 19 genes for dilated cardiomyopathy has recently been published.25

To further validate our HCM array for initial mutation screening in clinical practice, we analysed 122 patients with HCM and obtained a total mutation detection rate of 31% (38/122), which is lower than the mean total mutation detection rate of 42.4% (range13.3–60.9%).26 One explanation of the lower detection rate is that a number of variants initially reported as disease-causing mutations are now thought to be common SNPs. We found four such variants in 13 patients. The absence of pathogenic mutations in the remaining probands may be due to phenocopies of HCM (eg, Danon disease or Anderson-Fabry disease), mutations in a gene or sequence not tiled on the HCM array, the involvement of additional, as yet unidentified, genes for HCM, or the inability of the resequencing arrays to detect insertions/deletions. This limitation of the resequencing array technology lowers our mutation detection rate, as insertions/deletions account for up to 13% of all HCM mutations. They occur mainly in MYBPC3 (32% of all mutations; HGMD), which actually harbour ∼30% of all known HCM mutations. Finally, mutations may be hidden by a region of hybridisation failure, which occurred in 1.6% of the sequences.

These results show that the HCM array is highly appropriate for first-tier mutation screening in clinical practice, mostly because of the considerable gain in time, labour and cost. The complete analysis can be performed within a few weeks. Several samples can be analysed simultaneously, and the actual cost per patient (currently US$1500) is considerably lower than for Sanger sequencing of 12 HCM genes. The analysis not only identifies the causative mutation but also modifying SNPs, which may be of importance in improving the predictive value of the genotype. It is, furthermore, important to point out the small size of our cohort and to suggest that the true clinical sensitivity of the HCM array will emerge over time. In patients with no mutation identified by our HCM array, an eventual phenocopy should be considered, and MYBPC3 could be reanalysed by Sanger sequencing in order to unravel possible small insertions/deletions. Finally, targeted next-generation sequencing of HCM genes will become an affordable option.27

As mutations in the same disease gene can cause different types of primary cardiomyopathy, the HCM resequencing array may also be used for initial screening of other forms of cardiomyopathy, such as dilated cardiomyopathy (DCM), restrictive cardiomyopathy (RCM) or left ventricular non-compaction cardiomyopathy (LVNC). The HCM array covers nine DCM genes (ACTC1, MYH7, MYBPC3, TNNT2, TPM1, TNNC1, TNNI3, CSRP3, PLN), three RCM genes (MYH7, TNNT2, TNNI3) and three LVNC genes (MYH7, TNNT2 and ACTC1). So far, no published study has analysed the 12 genes present on the HCM array for patients with DCM, RCM or LVNC. Thus, it is not possible to derive an accurate sensitivity for these disorders, but the array can be proposed as a first, rapid and cheap mutation screening method for the index patients.

There is no doubt that next-generation sequencing will replace all large-scale sequencing platforms, as it already has been demonstrated in a few instances.27 28 However, these technologies are currently still expensive. For most diagnostic applications, targeted analysis of well-known pathogenic genes is at present preferred, as the interpretation of sequence variants is less complex than in whole exome sequencing, and it overcomes limitations in computational power as well as ethical concerns. The custom array sequencing technology has been validated for clinical application29 and provides a rapid, reasonably efficient and economically competitive method for molecular first-tier screening of Mendelian heterogeneous disorders such as HCM in clinical practice. It can currently be considered the method of choice until next-generation sequencing is routinely implemented in diagnostic laboratories.


We acknowledge P Descombes, C Barraclough, D Chollet, M Docquier from the Genomic Platform of Frontiers in Genetics, University School of Medicine, Geneva, Switzerland for technical support, and L Gwanmesia for proof reading. This study was supported by grants from the University Hospitals of Geneva, the Novartis Foundation Switzerland, the Valentine Gerbex-Bourget Foundation of Geneva, ACLON Finance Foundation Geneva and the Swiss Heart Foundation (to SF, JLB), a research grant from the Charité-Universitätsmedizin Berlin (to AP) and by the Italian Ministry for Scientific and Technologic Research (PM).


View Abstract


  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the University Hospitals of Geneva.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.