Get started
Work

Evaluation of exome sequencing variation in undiagnosed ataxias: Table 1

Year: 2015
Type: letter
Abstract: Sir, It is with great interest that we read the work of Pyle et al. (2014) on the utility of exome sequencing in achieving molecular diagnosis in ataxia. In particular, we take notice of the higher rate of ‘confirmed pathogenic’ and ‘possible pathogenic’ variants identified in this work as compared to others and question the stringency with which these classifications are applied. As stated by Pyle et al. (2014), the 41% (9/22 families) confirmed pathogenic rate is more than 2-fold higher than the 18% (9/50 individuals) detection rate reported using large targeted gene panels (Németh et al., 2013). This was explained by not using a limited gene panel in the Pyle et al. (2014) study, and possibly deeper coverage. However, Fogel et al. (2014), applied whole-exome sequencing to ataxia and found a similarly low rate of pathogenic variants of 21% (16/76 individuals). In this letter we point out that while promising variants were presented by Pyle et al. (2014), many molecular diagnoses that were claimed were based on data that were not entirely convincing, and are likely false-positive. Claiming a variant as pathogenic has immediate implications for future clinical testing and counselling, as publications such as Pyle et al. (2014) are used as a peer-reviewed public record to evaluate if a clinical variant is disease-causing or not in future cases. Claiming a variant as confirmed pathogenic when the evidence is less than convincing may result in inaccurate diagnosis, inappropriate changes in medical treatment, inaccurate counselling in pre-natal and presymptomatic testing environments, or abortion of healthy foetuses. A recent perspective in Nature provides recommendations as to what evidence would be needed to claim variant causality (MacArthur et al., 2014). In agreement with these recommendations, we raise three types of concerns: Allele frequency of the variant of interest, or minor allele frequency (MAF), in the general population should be low according to public variant databases. Several variants reported by Pyle et al. (2014) have a MAF > 0.01 in the databases provided despite a stated MAF cut-off of 0.01. Predicted damaging potential of a variant (pathogenicity), even within established genes for ataxia, needs to be carefully assessed. Multiple independent predictive and experimental approaches should be utilized to corroborate predicted pathogenicity (MacArthur et al., 2014). Prior association to the presented phenotype (phenotype). Clinical presentations for several variants claimed to cause ataxia by Pyle et al. differ significantly from those reported previously. Multiple unrelated cases with similar presentation should be reported before assigning a gene as causative for a particular phenotype (MacArthur et al., 2014). In all, we raise questions for three families with ‘confirmed pathogenic’ and four families with ‘possibly pathogenic’ classification. Table 1 summarizes the variants reported by Pyle et al. (2014) that we find questionable, and the type of concerns. Concerns on variants reported by Pyle et al. (2014) in patients with ataxia P5 P6 p.Ser527* C.Het/R p.Leu226Ile C.Het/R p.Met156Thr C.Het/R MAF: In-house MAF = 0.0157, ESP6500 MAF = 0.0115, 1000g MAF = 0.01 Phenotype:NPC1 typically more severe phenotype than the cases reported in this paper p.Pro237Ser C.Het/R MAF: In-house MAF = 0.0052, ESP6500 MAF = 0.0153, 1000g MAF = 0.0064 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 p.Arg454Gln Het/D MAF: In-house MAF not reported, ESP6500 MAF = 0.0003, 1000g MAF not reported Pathogenicity: Tolerated/neutral by SIFT and LRT Phenotype: Gene associated with childhood onset episodic ataxia, cases reported are adult onset progressive p.Lys193Gln C.Het/R Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: Variant associated with hearing loss (Cryns et al., 2003) p.Arg456His C.Het/R MAF: In-house MAF = 0.0402, ESP6500 MAF = 0.0467, 1000g MAF = 0.0600 Phenotype: Variant associated with diabetes (Cryns et al., 2003) c.-66A>G C.Het/R MAF: In-house MAF = 0.0122, ESP6500 MAF not reported, 1000g MAF = 0.0046 Pathogenicity: Splice site not molecularly confirmed Phenotype: Gene associated with infantile mitochondrial encephalopathy p.Lys50Arg C.Het/R c.805-2A>G Het/D Pathogenicity: Splice site not molecularly confirmed, other transcripts use alternate downstream splice site Phenotype: Gene associated with hereditary spastic paraplegia p.Thr665Ile C.Het/R MAF: In-house MAF = 0.0717, ESP6500 MAF = 0.0554, 1000g MAF = 0.0700 Phenotype: Gene reported in hereditary sensory and autonomic neuropathy p.Thr1091Ile C.Het/R p.Arg273Gln Hem/D MAF: In-house MAF = 0.0105, ESP6500 MAF = 0.0027, 1000g MAF = 0.0018 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: X-linked, patient is male with two affected children suggesting autosomal dominant inheritance p.Trp748Ser Het/D p.Glu1143Gly Het/D MAF: In-house MAF = 0.0262, ESP6500 MAF = 0.0275, 1000g MAF = 0.0200 Phenotype: POLG associated with recessive inherited ataxia. Variant reported as common haplotype with p.Trp748Ser (Van Goethem et al., 2004; Hakonen et al., 2005) p.Leu432Val Het/D P5 P6 p.Ser527* C.Het/R p.Leu226Ile C.Het/R p.Met156Thr C.Het/R MAF: In-house MAF = 0.0157, ESP6500 MAF = 0.0115, 1000g MAF = 0.01 Phenotype:NPC1 typically more severe phenotype than the cases reported in this paper p.Pro237Ser C.Het/R MAF: In-house MAF = 0.0052, ESP6500 MAF = 0.0153, 1000g MAF = 0.0064 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 p.Arg454Gln Het/D MAF: In-house MAF not reported, ESP6500 MAF = 0.0003, 1000g MAF not reported Pathogenicity: Tolerated/neutral by SIFT and LRT Phenotype: Gene associated with childhood onset episodic ataxia, cases reported are adult onset progressive p.Lys193Gln C.Het/R Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: Variant associated with hearing loss (Cryns et al., 2003) p.Arg456His C.Het/R MAF: In-house MAF = 0.0402, ESP6500 MAF = 0.0467, 1000g MAF = 0.0600 Phenotype: Variant associated with diabetes (Cryns et al., 2003) c.-66A>G C.Het/R MAF: In-house MAF = 0.0122, ESP6500 MAF not reported, 1000g MAF = 0.0046 Pathogenicity: Splice site not molecularly confirmed Phenotype: Gene associated with infantile mitochondrial encephalopathy p.Lys50Arg C.Het/R c.805-2A>G Het/D Pathogenicity: Splice site not molecularly confirmed, other transcripts use alternate downstream splice site Phenotype: Gene associated with hereditary spastic paraplegia p.Thr665Ile C.Het/R MAF: In-house MAF = 0.0717, ESP6500 MAF = 0.0554, 1000g MAF = 0.0700 Phenotype: Gene reported in hereditary sensory and autonomic neuropathy p.Thr1091Ile C.Het/R p.Arg273Gln Hem/D MAF: In-house MAF = 0.0105, ESP6500 MAF = 0.0027, 1000g MAF = 0.0018 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: X-linked, patient is male with two affected children suggesting autosomal dominant inheritance p.Trp748Ser Het/D p.Glu1143Gly Het/D MAF: In-house MAF = 0.0262, ESP6500 MAF = 0.0275, 1000g MAF = 0.0200 Phenotype: POLG associated with recessive inherited ataxia. Variant reported as common haplotype with p.Trp748Ser (Van Goethem et al., 2004; Hakonen et al., 2005) p.Leu432Val Het/D C.Het = compound heterozygous; D = dominant; Hem = hemizygous; Het = heterozygous; R = recessive. Concerns on variants reported by Pyle et al. (2014) in patients with ataxia P5 P6 p.Ser527* C.Het/R p.Leu226Ile C.Het/R p.Met156Thr C.Het/R MAF: In-house MAF = 0.0157, ESP6500 MAF = 0.0115, 1000g MAF = 0.01 Phenotype:NPC1 typically more severe phenotype than the cases reported in this paper p.Pro237Ser C.Het/R MAF: In-house MAF = 0.0052, ESP6500 MAF = 0.0153, 1000g MAF = 0.0064 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 p.Arg454Gln Het/D MAF: In-house MAF not reported, ESP6500 MAF = 0.0003, 1000g MAF not reported Pathogenicity: Tolerated/neutral by SIFT and LRT Phenotype: Gene associated with childhood onset episodic ataxia, cases reported are adult onset progressive p.Lys193Gln C.Het/R Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: Variant associated with hearing loss (Cryns et al., 2003) p.Arg456His C.Het/R MAF: In-house MAF = 0.0402, ESP6500 MAF = 0.0467, 1000g MAF = 0.0600 Phenotype: Variant associated with diabetes (Cryns et al., 2003) c.-66A>G C.Het/R MAF: In-house MAF = 0.0122, ESP6500 MAF not reported, 1000g MAF = 0.0046 Pathogenicity: Splice site not molecularly confirmed Phenotype: Gene associated with infantile mitochondrial encephalopathy p.Lys50Arg C.Het/R c.805-2A>G Het/D Pathogenicity: Splice site not molecularly confirmed, other transcripts use alternate downstream splice site Phenotype: Gene associated with hereditary spastic paraplegia p.Thr665Ile C.Het/R MAF: In-house MAF = 0.0717, ESP6500 MAF = 0.0554, 1000g MAF = 0.0700 Phenotype: Gene reported in hereditary sensory and autonomic neuropathy p.Thr1091Ile C.Het/R p.Arg273Gln Hem/D MAF: In-house MAF = 0.0105, ESP6500 MAF = 0.0027, 1000g MAF = 0.0018 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: X-linked, patient is male with two affected children suggesting autosomal dominant inheritance p.Trp748Ser Het/D p.Glu1143Gly Het/D MAF: In-house MAF = 0.0262, ESP6500 MAF = 0.0275, 1000g MAF = 0.0200 Phenotype: POLG associated with recessive inherited ataxia. Variant reported as common haplotype with p.Trp748Ser (Van Goethem et al., 2004; Hakonen et al., 2005) p.Leu432Val Het/D P5 P6 p.Ser527* C.Het/R p.Leu226Ile C.Het/R p.Met156Thr C.Het/R MAF: In-house MAF = 0.0157, ESP6500 MAF = 0.0115, 1000g MAF = 0.01 Phenotype:NPC1 typically more severe phenotype than the cases reported in this paper p.Pro237Ser C.Het/R MAF: In-house MAF = 0.0052, ESP6500 MAF = 0.0153, 1000g MAF = 0.0064 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 p.Arg454Gln Het/D MAF: In-house MAF not reported, ESP6500 MAF = 0.0003, 1000g MAF not reported Pathogenicity: Tolerated/neutral by SIFT and LRT Phenotype: Gene associated with childhood onset episodic ataxia, cases reported are adult onset progressive p.Lys193Gln C.Het/R Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: Variant associated with hearing loss (Cryns et al., 2003) p.Arg456His C.Het/R MAF: In-house MAF = 0.0402, ESP6500 MAF = 0.0467, 1000g MAF = 0.0600 Phenotype: Variant associated with diabetes (Cryns et al., 2003) c.-66A>G C.Het/R MAF: In-house MAF = 0.0122, ESP6500 MAF not reported, 1000g MAF = 0.0046 Pathogenicity: Splice site not molecularly confirmed Phenotype: Gene associated with infantile mitochondrial encephalopathy p.Lys50Arg C.Het/R c.805-2A>G Het/D Pathogenicity: Splice site not molecularly confirmed, other transcripts use alternate downstream splice site Phenotype: Gene associated with hereditary spastic paraplegia p.Thr665Ile C.Het/R MAF: In-house MAF = 0.0717, ESP6500 MAF = 0.0554, 1000g MAF = 0.0700 Phenotype: Gene reported in hereditary sensory and autonomic neuropathy p.Thr1091Ile C.Het/R p.Arg273Gln Hem/D MAF: In-house MAF = 0.0105, ESP6500 MAF = 0.0027, 1000g MAF = 0.0018 Pathogenicity: Tolerated/benign by SIFT and PolyPhen-2 Phenotype: X-linked, patient is male with two affected children suggesting autosomal dominant inheritance p.Trp748Ser Het/D p.Glu1143Gly Het/D MAF: In-house MAF = 0.0262, ESP6500 MAF = 0.0275, 1000g MAF = 0.0200 Phenotype: POLG associated with recessive inherited ataxia. Variant reported as common haplotype with p.Trp748Ser (Van Goethem et al., 2004; Hakonen et al., 2005) p.Leu432Val Het/D C.Het = compound heterozygous; D = dominant; Hem = hemizygous; Het = heterozygous; R = recessive. Due to the concerns presented in Table 1, these variants warrant more cautious interpretation. Additional detailed concerns for a select subset of cases are as follows: Pyle et al. (2014) report molecular confirmation of these variants by oxysterol analysis. There is a striking difference in values between the two siblings, with Patient P15 borderline normal for two measurements and normal for one measurement; and Patient P16 showing elevated results for all three measurements. No P-values are given and it is unclear if both patients are statistically outside of normal ranges. Other medical conditions can cause or have been associated with elevated oxysterol results, such as neuronal ceroid lipofuscinosis and Gaucher disease (Porter et al., 2010), hepatitis C infection (Arciello et al., 2012) and hyperlipidaemia (Arca et al., 2007). These alternative explanations should be excluded before making a diagnosis. Additional sequencing of unaffected siblings revealed that the siblings are not carriers of either variant and did not confirm whether these two mutations are compound heterozygous mutations (trans) or double mutants (cis). Pyle et al. (2014) refers to the variants reported in WFS1 as known mutations that have been previously described in patients with Wolfram syndrome, citing Cryns et al. (2003). However, these two mutations are not associated with Wolfram syndrome (Cryns et al., 2003). Mutations in WFS1 can cause Wolfram syndrome, primarily characterized by juvenile diabetes and early onset optic atrophy, but neurodegenerative features such as ataxia can develop in many cases. Mutations in WFS1 are also associated with type 1 and type 2 diabetes and deafness. Cryns et al. (2003), the Human Gene Mutation Database (Stenson et al., 2014), along with other researchers, have noted an association between Arg456His and development of diabetes and between Lys193Gln and deafness. Neither of these variants has been previously associated with neurodegenerative symptoms or ataxia. Pyle et al. (2014) used splicing prediction programs and found both a predicted gain and loss of an acceptor site. c.-66A>G sits in exon 1 in the 5’UTR of FASTKD2 near the splice donor site that does not appear to be impacted. The splice acceptor site at intron 1/exon 2 also does not appear to be impacted. A heterozygous splice site variant in ZFYVE27 is reported in a dominant family (Pyle et al., 2014). The Exome Aggregation Consortium (ExAC) reports 11 different splice, frameshift and stop gain variants in ZFYVE27, suggesting that haploinsufficiency alone isn’t likely to cause disease (ExAC, 2015). The reported splice site is not utilized in all transcripts for ZFYVE27, an additional splice site exists five amino acids further downstream, such that impact of possible differential splicing in the patients is unclear. With these considerations the ‘confirmed pathogenic’ rate drops to 6/22 families (27.2%), similar to the 21% found by other exome data (Fogel et al., 2014). These six families include the two childhood onset SACS variants families, the family with a previously reported KCNC3 variant, the family with previously reported SPG7 variants, and the two families with TUBB4A variants. We understand the importance of discovering a molecular diagnosis in currently undiagnosed cases of ataxia. It is incredibly frustrating for patients and physicians when there is no definitive explanation for a patient’s symptoms. Reports of variants of interest are helpful to the scientific community to explore if other researchers have found the same damaging variant in similar phenotypes. However, when additional molecular evidence or other affected individuals do not exist, more caution should be exercised before claiming a particular variant as causative, as such claims may lead to false clinical tests in the future. Research is funded by the National Institutes of Health (NIH) grant R01-NS078560.
Source: Brain
DOI: 10.1093/brain/awv087
Language: English
Subfield: Genetics
Open Access status: green
OpenAlex ID: W1964594915
Cites: 11
Cited by: 4
Related to: 10