SEARCH

SEARCH BY CITATION

Keywords:

  • NGS epilepsy panel;
  • Mutation screening;
  • Genetic diagnostics;
  • Massive parallel sequencing

Summary

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

Purpose:  Epilepsies have a highly heterogeneous background with a strong genetic contribution. The variety of unspecific and overlapping syndromic and nonsyndromic phenotypes often hampers a clear clinical diagnosis and prevents straightforward genetic testing. Knowing the genetic basis of a patient’s epilepsy can be valuable not only for diagnosis but also for guiding treatment and estimating recurrence risks.

Methods:  To overcome these diagnostic restrictions, we composed a panel of genes for Next Generation Sequencing containing the most relevant epilepsy genes and covering the most relevant epilepsy phenotypes known so far. With this method, 265 genes were analyzed per patient in a single step. We evaluated this panel on a pilot cohort of 33 index patients with concise epilepsy phenotypes or with a severe but unspecific seizure disorder covering both sporadic and familial cases.

Key Findings:  We identified presumed disease-causing mutations in 16 of 33 patients comprising sequence alterations in frequently as well as in less commonly affected genes. The detected aberrations encompassed known and unknown point mutations (SCN1A p.R222X, p. E289V, p.379R, p.R393H; SCN2A p.V208E; STXBP1 p.R122X; KCNJ10 p.L68P, p.I129V; KCTD7 p.L108M; KCNQ3 p.P574S; ARHGEF9 p.R290H; SMS p.F58L; TPP1 p.Q278R, p.Q422H; MFSD8 p.T294K), a putative splice site mutation (SCN1A c.693A> p.T/P231P) and small deletions (SCN1A p.F1330Lfs3X [1 bp]; MFSD8 p.A138Dfs10X [7 bp]). All mutations have been confirmed by conventional Sanger sequencing and, where possible, validated by parental testing and segregation analysis. In three patients with either Dravet syndrome or myoclonic epilepsy, we detected SCN1A mutations (p.R222X, p.P231P, p.R393H), even though other laboratories had previously excluded aberrations of this gene by Sanger sequencing or high-resolution melting analysis.

Significance:  We have developed a fast and cost-efficient diagnostic screening method to analyze the genetic basis of epilepsies. We were able to detect mutations in patients with clear and with unspecific epilepsy phenotypes, to uncover the genetic basis of many so far unresolved cases with epilepsy including mutation detection in cases in which previous conventional methods yielded falsely negative results. Our approach thus proved to be a powerful diagnostic tool that may contribute to collecting information on both common and unknown epileptic disorders and in delineating associated phenotypes of less frequently mutated genes.

Epilepsy is a common and highly heterogeneous neurologic disorder. Especially the pediatric epilepsies have a very diverse background, and seizures may often represent only one aspect of an underlying complex syndrome (Muthugovindan & Hartman, 2010).

Among hereditary factors known to account for epilepsy, all kinds of genetic aberrations have been described, for example, monogenic, complex, mitochondrial, chromosomal, and imprinting, (Claes et al., 2001; Weber & Lerche, 2008; Baulac & Baulac, 2010; Klassen et al., 2011; Mulley & Mefford, 2011) However, in the majority of cases seen in clinical practice the underlying genetic aberrations remain elusive. This may, of course, be due to unidentified epilepsy genes as well as complex inheritance of a subset of pathogenic or predisposing factors. However, it seems plausible that numerous patients with an epilepsy phenotype not suggestive of a specific genetic disorder may still have aberrations in known but rarely mutated genes.

Identifying the genetic background of a patient’s epilepsy phenotype may help in counseling the patient and relatives concerning a possibly elevated recurrence risk. The definitive diagnosis is important to avoid an often endless series of laborious, stressful, and expensive diagnostic procedures. In some cases it may also allow the prediction of future development or even guide antiepileptic pharmacotherapy, as for example by starting a ketogenic diet in glucose transporter type 1 (Glut1) deficiency or avoiding administration of Na+ channel blockers in Dravet syndrome (Brodie et al., 2011). Furthermore, detecting mutations in known epilepsy genes in patients with an uncommon or unspecific presentation of a seizure disorder may help to reveal the true phenotypic spectrum of the disorder.

Establishing a precise clinical diagnosis of a genetic disease with an unspecific epilepsy phenotype is a challenging task. Additional features such as accompanying disorders, patient and family history, laboratory parameters, dysmorphism, or malformation may help in finding the diagnosis only if they are specific. Screening for underlying genetic defects is often complicated by the multitude of putatively responsible genes, each with a low individual prevalence and mutation detection rate.

Mutation screening with conventional Sanger sequencing is considered the “gold standard” with respect to sequence quality. However, this method allows screening of only one gene at a time, exon by exon. In contrast, recently developed Next Generation Sequencing (NGS) techniques allow massive parallel sequencing of as many genes as desired. Here, the sequence quality depends on the minimum coverage at a base-pair position and the optimization of enrichment procedures of the genes of interest (target sequence) (Mamanova et al., 2010; Metzker, 2010). Targeted NGS seems particularly appropriate to screen for mutations in disorders with a highly heterogeneous genetic background, such as cardiomyopathies, neuromuscular diseases, retinopathies, and not least, epilepsies.

The power in detecting mutations in positive controls using an NGS panel of genes for ataxia and for congenital disorders of glycosylation has recently been shown (Hoischen et al., 2010; Jones et al., 2011). Shearer et al. (2010) successfully tested an NGS panel containing 54 genes related to hearing loss and revealed mutations in five of six patients.

We developed an NGS panel of genes comprising the most relevant epilepsy genes and epilepsy phenotypes (n = 265 genes). We screened a pilot cohort of 33 index patients representing both sporadic and familial cases with either concise epilepsy phenotypes or severe and often unspecific epileptic disorders, often with additional symptoms such as myoclonus or mental retardation. We aim to demonstrate how this technique may facilitate and accelerate the detection of the underlying genetic defect of a patient’s epilepsy syndrome.

Materials and Methods

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

Patients

We collected a cohort of 33 index cases with very different epilepsy phenotypes from various hospitals and clinical practices in Germany and Switzerland. The patients were selected randomly. A few patients had been screened previously for mutations in selected genes with conventional techniques. However, none of the patients had hitherto known genetic aberrations.

Written informed consent for genetic testing according to the respective national law was obtained from all tested individuals or their respective legal guardians prior to analysis.

Epilepsy panel

We chose 265 genes that were known to be involved in monogenic disorders including epilepsy as a phenotypic feature according to the Online Mendelian Inheritance in Man (OMIM) database (http://www.ncbi.nlm.nih.gov/omim). This panel of genes covers exonic regions as well as exon–intron boundaries of most of currently known epilepsy genes.

We used a customized target in solution enrichment followed by Next Generation Sequencing to screen for mutations in the 265 candidate genes (875,678 bp in total). Briefly, we fragmented the human genome, enriched for the coding regions of genes by using complementary highly specific RNA baits (Agilent Sure Select Target Enrichment, Agilent, Santa Clara, CA, U.S.A.) and then sequenced consecutively on a SOLiD 4 platform (Life Technologies, Carlsbad, CA, U.S.A.). We sequenced two patients on 1/8 (1 Oct) of a full slide. In total we sequenced 33 patients on 16.5 Octs. The hereby-generated 50 bp reads were mapped to the human genome and analyzed using the BioScope software versions 1.2. (Life Technologies). BioScope processes the reads with a Blast-like mapping algorithm using color codes. The single nucleotide variant (SNV) calling is implemented with a frequentist algorithm on high coverage positions or a Bayesian algorithm. Both algorithms fully support dicolors.

The detected variants were evaluated for their context to the individual phenotype (according to OMIM). All putatively causative mutations (Table 2) were validated by conventional Sanger sequencing. To further determine the impact of novel variants, we tested for segregation in families if possible. In all cases, we aimed to have complete SNV calling of all genes relevant for the patient’s phenotype. Hence, we resequenced all underrepresented regions (coverage ≤10 SOLiD reads) by Sanger in all relevant genes.

We tried to estimate the putative pathogenic effect of every suspected mutation with conventional and freely available online tools, such as Polyphen2 (http://genetics.bwh.harvard.edu/pph2/), MutationTaster (http://www.mutationtaster.org/MutationTaster/documentation.html) and Human Splicing Finder (http://www.umd.be/HSF/) (Desmet et al., 2009; Adzhubei et al., 2010; Schwarz et al., 2010).

Results

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

Patients

We sequenced the DNA of 33 patients with different epileptic syndromes (Table 2). Eight of these 33 patients had a well-defined phenotype, as for example Dravet syndrome, ceroid lipofuscinosis (CLN), or early infantile epileptic encephalopathies (EIEE), suggesting one or more candidate genes (subcohort A: 208, 214, 218, 239, 515, 520, 561, 688). Ten of 33 patients (subcohort B: 109, 219, 231, 264, 280, 310, 521, 525, 563, 802) were assigned to phenotypes for which the genetic background is largely unknown, such as rolandic epilepsy, myoclonic–astatic epilepsy, Lennox-Gastaut syndrome, or SCN1A-negative Dravet syndrome. The remaining 15 patients (subcohort C: 210, 211, 238, 243, 274, 303, 335, 345, 514, 517, 519, 524, 526, 555, 942) had unspecific manifestations of seizures with or without mental retardation. Table 2 provides an overview of symptoms, clinical diagnosis, and further affected family members of the 33 patients included in this study. The ages of onset were reported as available but were often unknown.

Ten patients had a positive family history regarding their individual disorder. The remaining 23 patients were sporadic cases.

Epilepsy panel

We chose 265 genes associated with epileptic seizures to create a comprehensive diagnostic tool with a diversified spectrum of application. To facilitate data analysis we further subdivided the 265 genes into 18 virtual subpanels according to clinically relevant groups of epilepsy phenotypes (Table 1). Some of the subpanels are overlapping to a certain extent due to phenotypic variability associated with certain epilepsy genes. This rough grouping of genes is not exhaustive but allows a prioritization of diagnostically important genes in an individual.

Table 1.   Assignment of 265 genes within the NGS epilepsy panel to different, partly overlapping subpanels according to rough phenotypic groups
Subpanel no.Subpanel/phenotypeNo. of genesGenes
  1. Genes in bold have been included in the update of the NGS epilepsy panel (version 2, containing 323 genes).

1Generalized/myoclonic/absence epilepsies, febrile seizures37 ALDH7A1 BRD2 CACNA1A CACNA1H CACNB4 CASR CHRNA2 CHRNA4 CHRNB2 CLCN2 CSTB EFHC1 EPM2A GABRA1 GABRB3 GABRD GABRG2 GPR98 GRIN2A GRIN2B KCNMA1 KCNQ2 KCNQ3 KCTD7 MBD5 ME2 NHLRC1 PCDH19 PRICKLE1 PRICKLE2 SCARB2 SCN1A SCN1B SCN2A SCN9A SLC2A1 TBC1D24
2Early infantile epileptic encephalopathies30 ARHGEF9 ARX CDKL5 CNTNAP2 FOXG1 GABRG2 GRIN2A GRIN2B MAPK10 MECP2 NRXN1 PCDH19 PNKP RNASEH2A RNASEH2B RNASEH2C SAMHD1 SCN1A SCN1B SCN2A SCN9A SLC25A22 SLC2A1 SLC9A6 SPTAN1 STXBP1 TCF4 TREX1 UBE3A ZEB2
3Neuronal migration disorders31 ARFGEF2 ARX COL18A1 COL4A1 CPT2 DCX EMX2 EOMES FGFR3 FKRP FKTN FLNA GPR56 LAMA2 LARGE PAFAH1B1 PAX6 PEX7 POMGNT1 POMT1 POMT2 PQBP1 RAB3GAP RELN SNAP29 SRPX2 TUBA1A TUBA8 TUBB2B VDAC1 WDR62
4Severe microcephaly/pontocerebellar hypoplasia22 ASPM ATR BUB1B CASK CDK5RAP2 CENPJ CEP152 LIG4 MCPH1 MED17 NHEJ1 PCNT PNKP PQBP1 RARS2 SLC25A19 STIL TSEN2 TSEN34 TSEN54 VRK1 WDR62
5Joubert syndrome and related disorders10 AHI1 ARL13B CC2D2A CEP290 CXORF5 INPP5E NPHP1 RPGRIP1L TMEM216 TMEM67
6Holoprosencephaly8 FGF8 GLI2 GLI3 PTCH1 SHH SIX3 TGIF ZIC2
7Epilepsy in X-linked mental retardation25 ARHGEF9 ARX ATP6AP2 ATRX CASK CDKL5 CUL4B CXORF5 DCX FGD1 GPC3 GRIA3 HSD17B10 JARID1C OPHN1 PAK3 PCDH19 PHF6 PLP1 PQBP1 RAB39B SLC9A6 SMC1A SMS SRPX2 SYP
8Selected inborn errors of metabolism32 ABCC8 ACY1 ADSL ALDH4A1 ALDH5A1 ALDH7A1 ARG1 ATIC BTD CPT2 DPYD ETFA ETFB ETFDH FH FOLR1 GAMT GCDH GCSH GCST GLDC HPD L2HGDH LAMA2 MOCS1 MOCS2 PGK1 PRODH QDPR SLC25A15 SLC46A1 SUOX
9Selected mitochondrial disorders35 APTX ATPAF2 BCS1L C12ORF65 C8ORF38 CABC1 COQ2 COQ9 COX10 COX15 DLD GCSH GCST GLDC HSD17B10 LRPPRC NDUFA2 NDUFS1 NDUFS3 NDUFS4 NDUFS7 NDUFS8 NDUFV1 PC PDHA1 PDSS1 PDSS2 POLG RARS2 SCO2 SDHA SURF1 TACO1 TMEM70 VDAC1
10Neuronal ceroid lipofuscinosis8 CLN3 CLN5 CLN6 CLN8 CTSD MFSD8 PPT1 TPP1
11Lysosomal storage disorders29 AGA ARSA ARSB CTSA FUCA1 GALC GALNS GLB1 GNE GNPTAB GNPTG GNS GUSB HEXA HEXB HGSNAT HYAL1 IDS IDUA MCOLN1 NAGLU NEU1 NPC1 NPC2 PSAP SGSH SLC17A5 SMPD1 SUMF1
12Congenital disorders of glycolization23 ALG1 ALG12 ALG2 ALG3 ALG6 ALG8 ALG9 B4GALT1 COG1 COG7 COG8 DPAGT1 DPM1 DPM3 GCS1 MGAT2 MPDU1 MPI PMM2 RFT1 SLC35A1 SLC35C1 TMEM15
13Disorders of peroxisome biogenesis9 PEX1 PEX12 PEX14 PEX2 PEX26 PEX3 PEX5 PEX6 PEX7
14Leukodystrophies20 ARSA ASPA EIF2B1 EIF2B2 EIF2B3 EIF2B4 EIF2B5 GALC GFAP MLC1 NOTCH3 PLP1 PSAP RNASEH2A RNASEH2B RNASEH2C SAMHD1 SDHA SUMF1 TREX1
15Disorders of the Ras/MAPK pathway14 BRAF CBL HRAS KRAS MAP2K1 MAP2K2 MYST4 NF1 NRAS PTPN11 RAF1 SHOC2 SOS1 SPRED1
16Migraine6 ATP1A2 CACNA1A NOTCH3 POLG SCN1A SLC2A1
17Hyperekplexia5 ARHGEF9 GLRA1 GLRB GPHN SLC6A5
18Syndromic disorders with epilepsy and others29 A2BP1 ATP2A2 ATP6V0A2 CCDC88C CLCNKA CLCNKB COH1 KCNA1 KCNJ1 KCNJ10 KIAA1279 LBR LGI1 MLL2 NIPBL PANK2 PI12 PIGV PLA2G6 RAI1 SCN8A SETBP1 SLC4A10 SMC3 SYNGAP1 TBX1 TSC1 TSC2 VPS13A

At least 35 million reads were generated per patient. Each read had a length of about 50 bp. The alignment of all reads to the human genome led to an average coverage of 292 per base pair (maximum coverage >2,500). About 98.4% of mapped reads were on targeted regions.

Table 2 summarizes the results of the panel analysis in all 33 patients of our cohort. We excluded all variants that were not on target and then detected on average 326 variants within the 265 panel genes per patient (range 209–640). Furthermore, we did not consider intronic variants outside of common splice regions (±30 bp) as well as synonymous exonic variations outside the highly conserved acceptor and donor splice sites. On average we identified 144 nonsynonymous exonic variants or putative splice variants per patient (range 72–288). Forty-eight of these variants (range 8–190) were novel to the dbSNP 131 (http://www.ncbi.nlm.nih.gov/projects/SNP/).

Table 2.   Patients analyzed with the NGS epilepsy panel with an overview over the individual patient data (left), the amount of detected variants in each patient (middle) as well as the mutations being detected, their validation and refinement of diagnosis (right)
Subcohort/caseSexAge in yearsSuggested clinical diagnosisSymptoms of index patientSymptoms of relativesVariants on targetns/splice variantsNovel to dbSNP131Novel to dbCeGaTMutations cDNA (detected by NGS and confirmed by Sanger sequencing)Mutations proteinMutation known to literatureSegregation with phenotypeFinal diagnosis
  1. Subcohort A comprises patients with a clear clinical diagnosis with one or several underlying genes. Subcohort B represents patients with a clear clinical phenotype or a certain etiologic spectrum with usually unknown or highly heterogeneous genetic background. Subcohort C covers patients with unspecific or diffuse phenotypes as well as patients with poor phenotypic data.

  2. ADHD, attention-deficit/hyperactivity disorder; BFIS, benign familial infantile seizures; BPNH, bilateral periventricular nodular heterotopia; CLN, ceroid lipofuscinosis, neuronal; dn, de novo; DD, developmental delay; (EI)EE, (early infantile) epileptic encephalopathy; f, female; FS, febrile seizures; GEFS+, generalized epilepsy with febrile seizures plus; GTCS, generalized tonic–clonic seizures; IGE, idiopathic generalized epilepsy; m, male; na, not analyzed; NS, nonsynonymous; PME, progressive myoclonus epilepsy; RSE, recurrent status epilepticus; SESAME, seizures, sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance; SUDEP, sudden unexpected death in epilepsy; XLMR, X-linked mental retardation.

  3. a Not detected by NGS epilepsy panel version 1.

  4. b SCN1A mutation negative in previous testing by different labs.

A 208m17EIEEGTCS, DDNo3031262113STXBP1 c.364C>TSTXBP1 p.R122XNonaEIEE4
A 214f29BFISClusters of GTCS between month 3 and 13Brother, 2 sisters, daughter, niece (similarly affected)2241041810SCN2A c.623T>ASCN2A p.V208ENoYesBFIS3
A 218f13DravetGTCS, myoclonus, FS, hypotonia, ataxia, DDNo21390116SCN1A c.866A>TSCN1A p.E289VNodnDravet
A 239f5PMEProgressive EE, GTCS, myoclonus, ataxia, DDBrother (similarly affected)3721635213KCTD7 c.322C>A (homozygous)KCTD7 p.L108M (homozygous)NoYesEPM3
A 515f8DravetGTCS, myoclonus, RSE, absences, hypotonia, DDNo2571182816SCN1A c.3990delTSCN1A p.F1330Lfs3XNonaDravet
A 520f40BPNHGTCS, periventricular nodular heterotopiaNo2411143013FLNA c.676G>TaFLNA p.R226XaYes1npBPNH
A 561f3EIEEGTCS, partial seizures, FS, DDNo2471162516SCN1A c.1136T>GSCN1A p.M379RNonaDravet
A 688f5CLNGTCS, DD/regression, vision impairmentNo54428819052MFSD8 c.881C>A/c.411_417del7MFSD8 p.T294K/A138Dfs10XYes2/noYesCLN7
B 109m12IGEFamilial absence epilepsyBrother and sister (absences), aunt (GTCS in childhood)2429995     
B 219f23GEFS+GTCS, myoclonus, FSSister, father (similarly affected)25910184     
B 231f33Lennox–GastautNonprogressive EE, GTCS, movement disorder, DDNo3501544316     
B 264m11Rolandic epilepsy & DDRolandic epilepsy, ADHD, DD, autistic behaviorNo327140213KCNQ3 c.1720C>TKCNQ3 p.P574SYes3naRolandic epilepsy
B 280m10GEFS+GTCS, absences (onset 5 years), FS, delay of speechBrother, sister, aunt (similarly affected)424153539     
B 310m53XLMRGTCS, DD2 Brothers (similarly affected)2271143622ARHGEF9 c.869G>AARHGEF9 p.R290HNonaEIEE8
B 521f9DravetGTCS, myoclonus, status, atypical absences, FS, DDNo231102128SCN1A c.1178G>AbSCN1A p.R393HbYes4dnDravet
B 525m7DooseGTCS, myoclonus, DDNo253116227     
B 563m5Dravet-likeMyoclonic astatic seizures, DDBrother (similarly affected)55228116119TPP1 c.833A>G/c.1266G>CTPP1 p.Q278R/p.Q422HNo/yes5YesCLN2
B 802m22DravetGTCS, myoclonus, FS, DDNo64027617277SCN1A c.664C>TbSCN1A p.R222XbYes6naDravet
C 210f32UnclearGTCS, progressive EE, ataxia, spasticity, DDNo298127265     
C 211m33UnclearGTCS, ataxia, spasticity, DDSister (similarly affected)3001212711KCNJ10 c.203T>C/c.385A>GKCNJ10 p.L68P/I129VNo/noYesSESAME
C 238f15UnclearGTCS, DD, dysgenesis of left hippocampusNo296135197     
C 243m30Partial epilepsy & DDGTCS, absences, partial seizures, myoclonus, DDNo350140257     
C 274m32UnclearMyoclonic epilepsyNo491211889     
C 303f5UnclearGTCS, DD, behavioral disorderNo467182595     
C 335m17UnclearGTCS, absences, dyskinesia, DDSister, mother (similarly but milder affected)296119248     
C 345m28UnclearGTCS in infancyNo2791294124     
C 514f2UnclearGTCS, RSE, microcephaly, DD, frontal polymicrogyriaNo209892010     
C 517m6UnclearGTCS, myoclonus, ataxia, hypsarrhythmia, mild DDNo228982416SCN1A c.693A>TbSCN1A p.231PbNodnAtypical Dravet
C 519f8UnclearGTCS, myoclonus, hypotonia, spasticity, DDNo2581031913     
C 524m23IGESporadic GTCS (onset 20 years), BrugadaBrother, father (rare GTCS), uncle, granduncle (SUDEP)235111239     
C 526f38IGEGTCS, dissociative disorderNo47022412724     
C 555m10UnclearPartial epilepsy, DD, subependymal heterotopiaNo47323213833     
C 942m23UnclearGTCS, EE, DD, scoliosis, severe osteoporosisNo21272126SMS c.426T>ASMS p.F58LNodnSnyder–Robinson
Average     3261444815     

Finally, we compared the variants with our in-house database. At the time of study, this database contained the data of about 100 individuals investigated by whole-exome sequencing on the same platform. Referral reasons of these 100 individuals were deafness, neurodegenerative diseases, as well different types of cancers. With this comparison we aimed to identify sequence platform specific errors. Variants that appeared ≥2× in this database might represent sequencing artifacts or so far unreported single-nucleotide polymorphisms (SNPs) and their analysis was thus postponed. This led to a further decrease of the number of truly novel variants down to 15 per patient (range 3–77). However, in patients with no apparent mutation we still evaluated all variants not annotated in dbSNP independently of being annotated in our in-house database.

By the design of our panel analysis, we targeted all coding regions of the 265 genes as well as all intron–exon boundaries. However, certain regions did not achieve a satisfying coverage of ≥10 and needed resequencing by Sanger (one polymerase chain reaction [PCR] for each region) to complete SNV calling (Fig. 1B). In patients without obvious mutations, the amount of resequencing depended on the amount of genes that were considered as being relevant according to the patient’s individual phenotype. Together with the validation of putatively causative variants, this led to a resequencing effort of up to 30 PCRs per patient.

image

Figure 1.   In comparison to whole exome enrichment and sequencing (A), the coverage (gray histogram) is increased for the targeted genes of the NGS epilepsy panel version 1 (B). However, some genes (e.g., FLNA– exons in blue) still were not satisfyingly covered and needed further improvement in version 2 of the panel (C) to allow a complete SNV calling. In parallel, the extent of underrepresented regions (yellow bars representing a coverage <10) decreased significantly.

Download figure to PowerPoint

Identification of mutations

We identified disease-causing aberrations in 16 of 33 patients (48%). Table 2 gives an overview of all mutations detected in our study.

In all eight patients of subcohort A (Table 2), the previously suggested clinical diagnosis could be confirmed by detection of mutations within one of the candidate genes. This comprises two patients with classical Dravet syndrome (218, 515) having mutations in SCN1A (p.E289V and p.F1330Lfs3X), a female with biparietal nodular heterotopias (520) due to an FLNA mutation (p.R226X) as well as a family with benign familial infantile seizures (index 214) and a SCN2A mutation (p.V208E). In two patients with the heterogeneous phenotype of EIEE (208, 561), we confirmed the diagnosis by detecting the mutations p.R122X in STXBP1 and p.M379R in SCN1A. Furthermore, we revealed the compound heterozygous mutations p.T294K/p.A138Dfs10X in MFSD8 as the cause for a patient’s late-infantile CLN (688). Finally, we detected the homozygous missense mutation p.L108M in KCTD7 in patient 239 with progressive myoclonus epilepsy (PME) as well as in the similarly affected sibling (Fig. 2B). Aberrations in KCTD7 have been described in only one single family before (Van Bogaert et al., 2007). Our family has therefore also been included in a cohort (Kousi et al., 2012) describing several new cases with PME and homozygous or compound heterozygous mutations in KCTD7 leading to the suggestion that KCTD7 might contribute more frequently to PME than previously estimated.

image

Figure 2.   (A) The heterozygous mutation p.V208E in SCN2A segregates with the phenotype within the family of index Patient 214 (arrow). All affected individuals were reported to have had at least one cluster of generalized seizures during the first months of life. None had developmental delay or seizures later than 13 months of age, confirming the diagnosis of BFIS. (B) Patient 239 with progressive myoclonus epilepsy as well as his similarly affected sister represent the second family ever described with a KCTD7 mutation. Both individuals were homozygous for p.L108M, whereas the healthy brother and consanguineous parents were heterozygous. (C) Both twin brothers with late-infantile ceroid lipofuscinosis were compound heterozygous for the mutations p.Q278R and p.Q422H in TPP1. Both parents were heterozygous for one mutation each. (D) The unspecific syndromic phenotype of epilepsy, ataxia, and mental retardation in two similarly affected siblings was attributed to compound heterozygosity of p.L68P and p.I129V in KCNJ10, leading to the very rare diagnosis of SESAME syndrome. Only the patient’s mother was available for genetic testing and appeared to be heterozygous for p.L68P in KCNJ10 only.

Download figure to PowerPoint

Among the 10 patients of subcohort B with distinctive clinical phenotypes of unknown genetic cause, we detected the underlying mutations in five Patients. Those included two patients with so far “SCN1A-negative Dravet syndrome” (521, 802) analyzed by other laboratories previous to our study, in which we detected SCN1A mutations (p.R393H and p.R222X). In Patient 563 as well as in his dizygotic twin brother, both with developmental delay, myoclonic–astatic seizures, and few clinical similarities to Dravet syndrome, we were able to detect compound heterozygosity for p.Q278R and p.Q422H in TPP1 leading to the diagnosis of late infantile CLN (Fig. 2C). Consequently performed enzyme analysis as well as a magnetic resonance imaging (MRI) investigation further confirmed this diagnosis. Hence, genetic analysis was able to establish diagnosis even before and independent of the precise clinical evaluation. In a family with three adult brothers with seizures and mental retardation, we identified the hemizygous X chromosomal mutation p.R290H in ARHGEF9 in the index patient (310) leading to the diagnosis of EIEE8. Unfortunately, we were not able to perform segregation analysis in this family. Finally, we detected the mutation p.P574S in KCNQ3 in a patient with rolandic epilepsy and additional features, such as mild developmental delay and abnormal behavior. According to the literature, the pathologic effect of this mutation is unclear (Miceli et al., 2009). However, it has been detected in four individuals with rolandic epilepsy from two different families as well as in eight individuals in a cohort of 455 IGE patients (Neubauer et al., 2008). It was not detected in any of the 454 controls. In dbSNP, p.P574S in KCNQ3 has been reported as rs74582884 in 4 of 2,188 alleles of the 1,000 Genome project (“minor allele frequency/minor allele count” of 0.0018/4) and is interpreted as being “pathogenic.” However, in our view this has to be interpreted with caution, since the pathogenicity of this variant has not been convincingly shown so far.

Among the 15 cases of subcohort C with unspecific clinical phenotypes, we detected aberrations in three patients. In Patient 517 with generalized and myoclonic seizures (onset 14 months), hypsarrhythmic EEG pattern, mild ataxia, and mild developmental delay, we detected the heterozygous de novo mutation c.693A>T in SCN1A. This aberration does not alter the amino acid sequence (p.P231P). However, it replaces a purine with a pyrimidine at the second last position in exon 5. According to MutationTaster and Human Splicing Finder (Desmet et al., 2009; Adzhubei et al., 2010; Schwarz et al., 2010), this results in loss of a sequence motif at the donor splice site and is likely to disturb normal splicing. It may therefore be responsible for a clinical phenotype with overlap but still some differences to classical Dravet syndrome. Surprisingly, SCN1A mutations had previously been excluded in this patient by high-resolution melting analysis. Furthermore, in a pair of sibs with an unspecific syndromic phenotype of epilepsy, ataxia, and mental retardation (index 211) we detected the mutations p.L68P and p.I129V in KCNJ10 in a compound heterozygous state (Fig. 2D). This led to the diagnosis of SESAME syndrome (seizures, sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance). Fewer than 10 families have been reported since its original description (Scholl et al., 2009). Finally, we detected a mutation in SMS predicting the amino acid substitution p.F58L in a male adult with an unspecific phenotype of generalized seizures, mental retardation, and severe osteoporosis with spontaneous fractures leading to the diagnosis of X-linked mental retardation of Snyder-Robinson type. This is again a very rare disorder, as only three families with affected male individuals have been described to date.

Validation of mutations and segregation analysis

All mutations have been validated and confirmed with conventional Sanger sequencing. The 16 mutation-positive cases comprise five patients (of 10/50%) with a positive family history and 11 sporadic cases (of 23/48%). In four of the five familial cases (211, 214, 239, 563), blood samples of the patient’s parents or relatives were available. Segregation of the individual mutations with the phenotype was confirmed within all four families (Fig. 2).

In three of 11 sporadic cases (218, 517, 521), parental analysis confirmed a de novo origin of the respective SCN1A mutation, as has been found in 95% of patients with Dravet syndrome (Nabbout et al., 2003). One male patient (942) had a mutation in SMS that was excluded in his mother, confirming a de novo origin of an X-linked disorder. In one further case (688), biparental inheritance could be confirmed for the compound heterozygous MFSD8 mutations.

Among the remaining six sporadic cases for whom no parental blood samples were available, three patients (264, 520, 802) were identified as carrying genetic aberrations in KCNQ3, FLNA, and SCN1A that have been already described in the literature (Claes et al., 2001; Parrini et al., 2006; Neubauer et al., 2008; Miceli et al., 2009) (Table 2). The other three (208, 515, 561) carried novel mutations in SCN1A (one nonsense, one missense) or STXBP1 (one nonsense).

In summary, the mutational spectrum comprised six known and 13 unknown mutations. Among them are two small deletions (SCN1A p.F1330Lfs3X [1 bp]; MFSD8 p.A138Dfs10X [7 bp]).

Discussion

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

Patients

With a mutation detection rate of 48% (16/33) our approach proved to be an efficient tool in identifying the underlying genetic defects of patients with highly varying epileptic phenotypes.

The detection of mutations in all eight individuals of subcohort A (Table 2) proves that a precise clinical phenotyping already points to specific genes. However, when a particular phenotype can be due to mutations in several genes, direct screening with conventional methods is time- and cost-consuming. For example, late-infantile CLN was found to be due to compound heterozygous mutations in MFSD8 in patient 688. In a conventional testing strategy, MFSD8 would have probably been one of the last genes screened, because mutations in TPP1, PPT1, CTSD, CLN5, and CLN6 and maybe also CLN8 are considered to be more frequent compared to MFSD8 (Mole & Williams, 2010).

We detected mutations in 5 of 10 patients of subcohort B. In these patients, conventional sequencing approaches can be very challenging due to the multitude of putatively causative genes responsible for certain phenotypes. On the other hand, many distinct epilepsy phenotypes have a widely unclear etiology and their genetic defects often remain undiscovered (as for example in Lennox–Gastaut syndrome or SCN1A-negative Dravet syndrome). Since the epilepsy panel covers a wide spectrum of differential diagnoses in epilepsy phenotypes, the detection of mutations is significantly facilitated in comparison to a one by one conventional Sanger sequencing approach in these cases.

Finally, we detected mutations in 3 of 15 patients of subcohort C. In these cases, conventional sequencing approaches would most probably fail to reveal genetic aberrations as an unspecific clinical phenotype hampers the prediction of putatively causative genes. However, sequencing the epilepsy panel in this subcohort allowed the detection of mutations in patients with poor or diffuse phenotypic data.

Thus, in subcohort B, our results led to a refinement of the clinical diagnosis in many cases, whereas in subcohort C, the panel analysis was the first opportunity to enable a diagnosis in at least some of the patients.

Epilepsy panel

In comparison to whole exome sequencing (WES), the largely decreased number of genes within a disease-specific gene panel allows a significant increase of coverage on target sequences (Fig. 1). This helps to reliably detect disease-causing mutations as well as to prevent detection, validation, and interpretation of countless variants in genes not suspected to be related to a patient’s phenotype.

Of course, this method is only suited in diagnostics of monogenic disorders or those with at least one major gene effect. It is not suited for complex genetic disorders for which WES as a research tool has the advantage of covering all coding genes, although with less abundant coverage.

We identified the well-known de novo mutations p.R222X and p.R393H in SCN1A (Claes et al., 2001; Ohmori et al., 2006) in two patients with classical Dravet syndrome (521, 802), although records for both patients stated negative mutation testing for SCN1A with conventional Sanger sequencing by approved laboratories in 2003 and 2009. In a third case (517), SCN1A mutations had been excluded by high-resolution melting analysis in a different laboratory. Nevertheless, we revealed the de novo putative splice mutation p.P231P in SCN1A. After detecting the mutations by the NGS epilepsy panel, we clearly confirmed their presence with Sanger sequencing in all three cases. According to one of the previous labs, the failure in detecting the p.R393H mutation was abetted by a 2 bp deletion in the adjacent intron, representing a known polymorphism that led to a misalignment of the two alleles and caused misinterpretation of the sequence data. This alignment artifact is specific to Sanger sequencing analysis and generally does not occur after mapping of data of the epilepsy panel or any other NGS approach. Unfortunately, we were not able to reconstruct the conditions that led to the failure in detecting the p.R222X and the p.P231P mutations by the other two laboratories.

In Patient 520 with bilateral periventricular heterotopia, our panel approach failed to detect a causative mutation in the first step. Investigation of FLNA, the most suspicious gene in this disorder, revealed unsatisfying coverage of several exons impeding a complete SNV calling (Fig. 1B). Subsequent Sanger sequencing of underrepresented regions revealed the known mutation p.R226X in FLNA (Parrini et al., 2006). This case illustrates that sensitivity of the panel analysis is substantially dependent on a sufficient coverage of the targeted regions.

To further improve the future diagnostic strength of the NGS epilepsy panel, we consequently optimized the coverage of genes/areas that did not perform optimally in the first version of the panel. We abandoned enrichment of regions that were out of diagnostic focus (e.g., 5′ and 3′ UTRs) and used the spared capacity to increase coverage of underrepresented regions. In addition, we included more genes (323 in total, Table 1) that we thought to be useful to complete the list of clinical relevant epilepsy genes compared to the first version of the panel (containing 265 genes). Preliminary data from this second version of the panel, which are not included in this paper, already show a significant improvement. Genes with gaps or drops of coverage, such as FLNA, had a far better and more even distribution of coverage and now met diagnostic requirements (Fig. 1C). This also led to a significant decrease of the amount of Sanger resequencing.

In addition to the described mutations, we also detected numerous unclassified variants in every single patient—mainly in genes that we doubt to be correlated with the patient’s individual phenotype. Some of these variants represented false positives and sequencing artifacts that were not confirmed by conventional Sanger sequencing. Others do represent known and unknown rare SNPs. At present, we were not able to validate all these variants. Especially in those cases where we were able to detect pathogenic mutations, we tended to neglect a possible but improbable major effect of these variants on the patient’s phenotype. In cases so far unresolved, the multitude of unclassified variants remains a significant problem because validation of all such variants and unequivocal exclusion of their possible pathogenic influences remain elusive at present. Hence, massive parallel sequencing of panel genes should be considered as a screening tool to detect clear monogenic or oligogenic mutations. Rare variants may contribute as a disease modifier to the genetic background of a patient’s disorder, but at present we are unable to estimate such contribution. Large cohorts and/or families and sophisticated biostatistical evaluation will be necessary to unravel this important question in genetic diseases such as familial epilepsy syndromes.

Our mutation detection algorithm focuses on variants novel to dbSNP 131. However, later versions of this database and inclusion of other sources (such as data from the 1,000 Genome project) might help to further decrease the number of truly novel variants.

For the 17 patients for whom our method failed to detect definite or putative mutations, it remains unclear whether this was due to (1) the causative gene not being among the 265 genes of the first version of the panel or not being sufficiently covered; (2) the causative mutation not being on target (e.g., both nonexonic and non–splice-site); (3) the causative mutation being detected but not identified (e.g., masked as unclassified variant) or missed (e.g., due to too stringent filtering); or (4) polygenic inheritance, chromosomal aberrations, CNVs, or other issues/problems.

In some of those 17 unresolved patients, however, we detected “novel” variants that might possibly be causative for the individual phenotypes. Due to the unavailability of parental blood samples, the validation of these variants was often exclusively restricted to in silico calculations and failed to deliver convincing evidence for a substantial confirmation of pathogenicity. However, we cannot exclude putatively causative effects for all these variants (data not shown). Still, similar to routine conventional genetic testing, the availability of parental blood samples may often facilitate the evaluation of a novel variant.

With our study, we demonstrate the efficiency of detecting genetic aberrations in patients with epilepsy using an NGS panel approach. In several cases we detected mutations in genes that were uncommonly mutated and are thus not routinely screened in conventional genetic testing strategies using Sanger sequencing. With the epilepsy panel, the detection rate of genetic aberrations no longer depends on the prerequisite of a precise clinical diagnosis and consecutive prediction of putatively casual genes. It rather seems sufficient to categorize the patient’s epilepsy in a rough group of phenotypes (e.g., according to a subpanel) and to evaluate mainly the respective genes. With the panel approach it was possible to uncover diagnoses in cases with unspecific phenotypes or poor clinical data (e.g., 211, 310, 942). Of interest, some of the diagnoses revealed by the epilepsy panel represent extremely rare disorders (e.g., 211, 239, 310, 942). In patients with more than one gene being suspected to account for the phenotype, the epilepsy panel turned out to be superior to conventional Sanger sequencing in both processing time and analysis cost. The cost of this panel analysis will be comparable to sequencing 1–3 medium- to large-sized genes by conventional techniques.

The detection of SCN1A mutations in three apparently SCN1A-negative cases underlies the diagnostic strength of the epilepsy panel in comparison to conventional techniques. Hence, even in cases with a clear phenotype and only one gene in diagnostic focus, primary screening with the epilepsy panel is certainly worth consideration.

In summary, we feel that this newly developed disease-specific NGS panel of genes represents a powerful tool for uncovering underlying genetic defects and for establishing a definite diagnosis in patients with various epilepsy disorders, and it strongly contributes to a better understanding of genotype–phenotype correlations.

Acknowledgments

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

This study has been partially funded by the Novartis Stiftung für Biologisch-Medizinische Forschung (10B46), by the Swiss League Against Epilepsy (Forschungsförderpreis 2010), by the Bundesministerium für Bildung und Forschung in Germany (BMBF: NGFNplus, EMINet/01GS08123; Netzwerk Seltene Erkrankungen, IonNeurONet/01GM1105A, TP5 and TP7) and by the European Science Foundation (Eurocores programme EuroEPINOMICS) through the Deutsche Forschungsgemeinschaft (DFG/Le1030/11-1) and the Schweizer Nationalfond (SNF 32EP30_136042/1). We are grateful to the support through Alexander Sartori and Alain Rico (Life Technologies).

Disclosure

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References

YGW serves/has served on scientific advisory boards for UCB Pharma and received honoraria from UCB Pharma and Desitin for participation in workshops. HL serves on scientific advisory boards for Eisai, GSK, Pfizer, UCB, and Valeant; receives industry-funded travel costs from GSK, Pfizer, and UCB; receives honoraria for speaking engagement or educational activities from Desitin, Eisai, GSK, Pfizer, and UCB; and receives research support from Sanofi-Aventis and UCB. GK serves on scientific advisory boards for Eisai, GSK, and UCB and receives honoraria for speaking engagement or educational activities from Desitin, Eisai, GSK, and UCB. TD receives honoraria for speaking engagement or educational activities from UCB, GSK, Janssen-Cilag, Eisai, Desitin, and Novartis. The remaining authors have no potential conflicts of interest to disclose.

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

References

  1. Top of page
  2. Summary
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  • Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, Kondrashov AS, Sunyaev SR. (2010) A method and server for predicting damaging missense mutations. Nat Methods 7:248249.
  • Baulac S, Baulac M. (2010) Advances on the genetics of Mendelian idiopathic epilepsies. Clin Lab Med 30:911929.
  • Brodie MJ, Covanis A, Gil-Nagel A, Lerche H, Perucca E, Sills GJ, White HS. (2011) Antiepileptic drug therapy: does mechanism of action matter? Epilepsy Behav 21:331341.
  • Claes L, Del-Favero J, Ceulemans B, Lagae L, Van Broeckhoven C, De Jonghe P. (2001) De novo mutations in the sodium-channel gene SCN1A cause severe myoclonic epilepsy of infancy. Am J Hum Genet 68:13271332.
  • Desmet FO, Hamroun D, Lalande M, Collod-Beroud G, Claustres M, Beroud C. (2009) Human Splicing Finder: an online bioinformatics tool to predict splicing signals. Nucleic Acids Res 37:e67.
  • Hoischen A, Gilissen C, Arts P, Wieskamp N, van der Vliet W, Vermeer S, Steehouwer M, de Vries P, Meijer R, Seiqueros J, Knoers NV, Buckley MF, Scheffer H, Veltman JA. (2010) Massively parallel sequencing of ataxia genes after array-based enrichment. Hum Mutat 31:494499.
  • Jones MA, Bhide S, Chin E, Ng BG, Rhodenizer D, Zhang VW, Sun JJ, Tanner A, Freeze HH, Hegde MR. (2011) Targeted polymerase chain reaction-based enrichment and next generation sequencing for diagnostic testing of congenital disorders of glycosylation. Genet Med 13:921932.
  • Klassen T, Davis C, Goldman A, Burgess D, Chen T, Wheeler D, McPherson J, Bourquin T, Lewis L, Villasana D, Morgan M, Muzny D, Gibbs R, Noebels J. (2011) Exome sequencing of ion channel genes reveals complex profiles confounding personal risk assessment in epilepsy. Cell 145:10361048.
  • Kousi M, Siintola E, Dvorakova L, Vlaskova H, Turnbull J, Topcu M, Yuksel D, Gokben S, Minassian BA, Elleder M, Mole SE, Lehesjoki AE. (2009) Mutations in CLN7/MFSD8 are a common cause of variant late-infantile neuronal ceroid lipofuscinosis. Brain 132:810819.
  • Kousi M, Anttila V, Schulz A, Calafato S, Jakkula E, Riesch E, Myllykangas L, Kalimo H, Topcu M, Gokben S, Alehan F, Lemke JR, Alber M, Palotie A, Kopra O, Lehesjoki AE. (2009) Novel KCTD7 mutations establish association to progressive myoclonus epilepsy. J Med Genet. in press.
  • Mamanova L, Coffey AJ, Scott CE, Kozarewa I, Turner EH, Kumar A, Howard E, Shendure J, Turner DJ. (2010) Target-enrichment strategies for next-generation sequencing. Nat Methods 7:111118.
  • Metzker ML. (2010) Sequencing technologies – the next generation. Nat Rev Genet 11:3146.
  • Miceli F, Soldovieri MV, Lugli L, Bellini G, Ambrosino P, Migliore M, del Giudice EM, Ferrari F, Pascotto A, Taglialatela M. (2009) Neutralization of a unique, negatively-charged residue in the voltage sensor of K V 7.2 subunits in a sporadic case of benign familial neonatal seizures. Neurobiol Dis 34:501510.
  • Mole SE, Williams RE. (2010) Neuronal ceroid-lipofuscinoses. In Pagon RA, Bird TD, Dolan CR, Stephens K (Eds) GeneReviews [Internet]. University of Washington, Seattle, WA, pp. 2001.
  • Mulley JC, Mefford HC. (2011) Epilepsy and the new cytogenetics. Epilepsia 52:423432.
  • Muthugovindan D, Hartman AL. (2010) Pediatric epilepsy syndromes. Neurologist 16:223237.
  • Nabbout R, Gennaro E, Dalla Bernardina B, Dulac O, Madia F, Bertini E, Capovilla G, Chiron C, Cristofori G, Elia M, Fontana E, Gaggero R, Granata T, Guerrini R, Loi M, La Selva L, Lispi ML, Matricardi A, Romeo A, Tzolas V, Valseriati D, Veggiotti P, Vigevano F, Vallee L, Dagna Bricarelli F, Bianchi A, Zara F. (2003) Spectrum of SCN1A mutations in severe myoclonic epilepsy of infancy. Neurology 60:19611967.
  • Neubauer BA, Waldegger S, Heinzinger J, Hahn A, Kurlemann G, Fiedler B, Eberhard F, Muhle H, Stephani U, Garkisch S, Eeg-Olofsson O, Muller U, Sander T. (2008) KCNQ2 and KCNQ3 mutations contribute to different idiopathic epilepsy syndromes. Neurology 71:177183.
  • Ohmori I, Kahlig KM, Rhodes TH, Wang DW, George AL Jr. (2006) Nonfunctional SCN1A is common in severe myoclonic epilepsy of infancy. Epilepsia 47:16361642.
  • Parrini E, Ramazzotti A, Dobyns WB, Mei D, Moro F, Veggiotti P, Marini C, Brilstra EH, Dalla Bernardina B, Goodwin L, Bodell A, Jones MC, Nangeroni M, Palmeri S, Said E, Sander JW, Striano P, Takahashi Y, Van Maldergem L, Leonardi G, Wright M, Walsh CA, Guerrini R. (2006) Periventricular heterotopia: phenotypic heterogeneity and correlation with Filamin A mutations. Brain 129:18921906.
  • Scholl UI, Choi M, Liu T, Ramaekers VT, Hausler MG, Grimmer J, Tobe SW, Farhi A, Nelson-Williams C, Lifton RP. (2009) Seizures, sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance (SeSAME syndrome) caused by mutations in KCNJ10. Proc Natl Acad Sci U S A 106:58425847.
  • Schwarz JM, Rodelsperger C, Schuelke M, Seelow D. (2010) MutationTaster evaluates disease-causing potential of sequence alterations. Nat Methods 7:575576.
  • Shearer AE, DeLuca AP, Hildebrand MS, Taylor KR, Gurrola J 2nd, Scherer S, Scheetz TE, Smith RJ. (2010) Comprehensive genetic testing for hereditary hearing loss using massively parallel sequencing. Proc Natl Acad Sci U S A 107:2110421109.
  • Steinfeld R, Steinke HB, Isbrandt D, Kohlschutter A, Gartner J. (2004) Mutations in classical late infantile neuronal ceroid lipofuscinosis disrupt transport of tripeptidyl-peptidase I to lysosomes. Hum Mol Genet 13:24832491.
  • Van Bogaert P, Azizieh R, Desir J, Aeby A, De Meirleir L, Laes JF, Christiaens F, Abramowicz MJ. (2007) Mutation of a potassium channel-related gene in progressive myoclonic epilepsy. Ann Neurol 61:579586.
  • Weber YG, Lerche H. (2008) Genetic mechanisms in idiopathic epilepsies. Dev Med Child Neurol 50:648654.