Dissecting genetic architecture of rare dystonia: genetic, molecular and clinical insights

Background Dystonia is one of the most common movement disorders. To date, the genetic causes of dystonia in populations of European descent have been extensively studied. However, other populations, particularly those from the Middle East, have not been adequately studied. The purpose of this study is to discover the genetic basis of dystonia in a clinically and genetically well-characterised dystonia cohort from Turkey, which harbours poorly studied populations. Methods Exome sequencing analysis was performed in 42 Turkish dystonia families. Using co-expression network (CEN) analysis, identified candidate genes were interrogated for the networks including known dystonia-associated genes and genes further associated with the protein-protein interaction, animal model-based characteristics and clinical findings. Results We identified potentially disease-causing variants in the established dystonia genes (PRKRA, SGCE, KMT2B, SLC2A1, GCH1, THAP1, HPCA, TSPOAP1, AOPEP; n=11 families (26%)), in the uncommon forms of dystonia-associated genes (PCCB, CACNA1A, ALDH5A1, PRKN; n=4 families (10%)) and in the candidate genes prioritised based on the pathogenicity of the variants and CEN-based analyses (n=11 families (21%)). The diagnostic yield was found to be 36%. Several pathways and gene ontologies implicated in immune system, transcription, metabolic pathways, endosomal-lysosomal and neurodevelopmental mechanisms were over-represented in our CEN analysis. Conclusions Here, using a structured approach, we have characterised a clinically and genetically well-defined dystonia cohort from Turkey, where dystonia has not been widely studied, and provided an uncovered genetic basis, which will facilitate diagnostic dystonia research.


INTRODUCTION
Dystonia is one of the most common movement disorders and is characterised by sustained or intermittent muscle contractions that cause abnormal and often repetitive movements, postures or both. 1 Hereditary dystonias are a clinically and genetically heterogeneous group of disorders.To date, based on the nomenclature of Genetic Movement Disorders 2 and the recent publications, 3 4 ~60 inherited forms of dystonia have been reported.A small proportion of those belong to an isolated dystonia group (40%) that has been associated with the autosomal recessive (AR) form of dystonia, suggesting the scarcity, genetic complexity and heterogeneity of the disease as well as the limitations of existing approaches, which have focused on specific populations.
So far, dystonia has been widely studied in European populations, [5][6][7][8] and it has been scarcely investigated in Middle Eastern, Asian 9 10 and African populations.Turkey lies at the crossroads between Europe and Asia.Due to Turkey's geographical location, its population is a genetic mix of populations from the Balkans, Caucasus, Middle East, Central Asia and Europe. 11Studies on Turkish cohorts are also of great importance in dystonia research in view of the high degree of consanguinity observed in Turkey. 11n this study, our major purpose was to dissect the complex architecture of dystonia through the identification of rare genetic causes in a wellcharacterised study cohort from Turkey and to provide an uncovered genetic background of dystonia to aid in clinical-diagnostic settings.For this purpose, we have conducted a comprehensive

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ The genetic causes of dystonia have been extensively studied in European populations and different diagnostic yields ranging from 11.7% to 39.1% have been reported.

WHAT THIS STUDY ADDS
⇒ This is the first study that covers dystonia families from Turkey, the population of which is known to be a mixture of populations from the Middle East, the Caucasus, the Balkans, Central Asia and Europe.⇒ By identifying the genetic causes of dystonia in a well-defined cohort of 42 families, we provide a distinct overview of the genetic makeup of dystonia in Turkey, and identify candidate genes and related mechanisms.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
⇒ This study will (1) aid in clinical-diagnostic settings in Turkey and some other related populations, such as Balkan, Caucasus and Middle Eastern populations and (2) contribute to dystonia research in terms of identifying candidate disease mechanisms.

Neurogenetics
study using a wide-scale genetic analysis approach.Through identifying known and new causes of the disease, we have shown that the genetic background of dystonia is overly complex and needs to be investigated systematically in different populations.

MATERIALS AND METHODS Patient's characteristics and clinical approach
A total of 42 dystonia families, which are compatible with the AR mode of inheritance, including 46 affected and 47 nonaffected individuals, were recruited in Turkey for the study at the Movement Disorders Unit, Department of Neurology, Istanbul University, between 2012 and 2015.Fifteen of those families were included in this study as trios and three as quartets.
The families were included in the study based on the consensus definition of dystonia. 1All the affected members and available unaffected family members underwent a detailed clinical examination.Radiological and laboratory findings were obtained from all the recruited individuals when needed.Consanguinity was reported in 16 families (38%).The mean age at onset of the affected members was 21.3±18 years (range 1-68 years).
In terms of clinical characteristics, the majority of our cohort were categorised into childhood onset (n=18; 42.8%), generalised (n=21; 49.9%) and isolated (n=28; 66.6%) dystonia based on the age at onset, body distribution and other associated features.The clinical characteristics are summarised in table 1, the details of which are provided in online supplemental materials S1-S4.

Genetic analyses
A systematic genetic screening strategy was conducted to analyse 93 exome sequencing (ES) samples from 42 families (see online supplemental material S5 for the workflow).

Exome sequencing analysis
ES data were processed as described in the online supplemental materials S5 and S6.The variants in the vcf files were first filtered for the dystonia-associated genes based on the dystonia-associated gene list, which was developed through a systematic search on OMIM and SCAIView neuro 12 (for the details of the genes, see online supplemental material S18), and the clinical significance of the retrieved variants were interpreted using Intervar. 13The pathogenic or likely pathogenic variants based on the American College of Medical Genetics and Genomics (ACMG)/Association for Molecular Pathology (AMP) guidelines or variant of uncertain significance (VUS) variants in the dystonia-associated genes were kept when (1) the clinical presentation of the family is highly specific for the identified gene variant, (2) the variant is determined as deleterious based on at least three pathogenicity scores obtained from dbNSFP 14 using SIFT, 15 CADD (Combined Annotation Dependent Depletion)(Phred score ≥15), 16 Poly-Phen2, 17 MutationTaster 18 and the Mendelian Clinically Applicable Pathogenicity score 19 or Splicing Clinically Applicable Pathogenicity 20 score (for the splicing variants) and LIST (Local Identity and Shared Taxa) 21 (and (3) the variant is not present as a homozygous state in the control datasets including GnomAD 22 and Turkish (TR) Variome. 11o identify the most plausible variants in the candidate genes, the remaining variants were prioritised based on (1) at least two prediction scores determining pathogenic and (2) minor allele frequency (MAF) in the control datasets (no homozygotes).The candidate genes were further interrogated for the enrichment with the established dystonia genes in the dystonia-associated CEN modules, 23 (2) involvement in an interaction pattern in a GeneMANIA network generated using the genes in the dystonia-associated WGCNA modules and (3) possible dystoniaassociated findings (reported phenotypes in mouse models with established gene mutations) in the animal models (AM) using 'The Human-Mouse: Disease Connection' 24 and 'The International Mouse Phenotyping Consortium'. 25The candidate genes in the modules that are significantly enriched with the dystoniaassociated genes were prioritised as the most plausible candidate genes (online supplemental material S5).
All the retrieved variants were confirmed and co-segregation analysis was performed by Sanger sequencing using the Big Dye Terminator Cycle Sequencing Ready Reaction kit on an ABI 3730 automated sequencer.
The statistical analyses were performed as described in the online supplemental material S7.

CEN analysis
Dystonia is defined as a network disorder with involvement of a basal ganglia (BG)-cerebello-thalamo-cortical circuit. 26Therefore, using CEN analysis, we interrogated the BG and cerebellum (CRBL) regions (discovery data) of the brain obtained from the the Genotype-Tissue Expression Consortium (GTEx Consortium, 2016; V.7) 23 database and the putamen region (replication data) of the brain obtained from the United Kingdom Brain Expression Consortium (UKBEC) 27 database.We could analyse the caudate, nucleus accumbens and putamen regions of the BG as the discovery dataset, but only the putamen region as the replication dataset due to the unavailability of the datasets.In

Neurogenetics
the CEN analysis, the candidate genes were interrogated for the modules, which are statistically enriched for the known dystoniaassociated genes and involve our candidate genes as described in the online supplemental material S8.Fisher's exact test was used to determine the modules that are statistically enriched for the dystonia-associated genes.
The gene set enrichment and over-representation (GSEO) and protein-protein interaction (PPI) analyses are described in the online supplemental material S8.

Overview of the genetic findings
Our ES analysis in 42 dystonia families allowed us to identify (1) pathogenic, likely pathogenic and VUS variants in the established and uncommonly dystonia-associated genes (n=15 (n=3 published elsewhere 28 29 ); 36%) and (2) variants in the candidate genes prioritised based on the CEN-based evidence (n=9; 21%) (figure 1A).There is no statistical association between the mutation status and either age at onset or associated features.
All the pedigrees corresponding to the reported families are illustrated in the online supplemental materials S9, S10.

Genetic findings in the dystonia-associated genes
Our ES approach revealed pathogenic, likely pathogenic and VUS variants in the established dystonia genes (PRKRA, SGCE, KMT2B, SLC2A1, GCH1, THAP1, HPCA, TSPOAP1 (BZRAP1), AOPEP (C9orf3)) as well as in the uncommonly dystoniaassociated genes (PCCB, CACNA1A, ALDH5A1, PRKN) (table 2).Two variants were classified as VUS in the PRKRA (n=1 family) and SGCE (n=1 family) genes.However, because additional analysis (eg, structural variant analysis) did not reveal any other plausible variants and because the pathogenicity scores and the MAF-based findings supported the high possibility of pathogenicity, those variants were also listed in this study.
The most frequently identified variants in the established dystonia genes in the diagnostic group were missense (n=12/17; 70%), followed by two stop-gain, two frameshift and one splicing variants (table 2, figure 1C).
Our analysis also uncovered pathogenic or likely pathogenic variants in uncommonly dystonia-associated genes (PCCB, CACNA1A, ALDH5A1, PRKN) in four families where dystonia has been rarely observed in the spectrum of the clinical characteristics so far.
We also identified a homozygous truncating variant in the AOPEP (C9orf3) gene, which has been recently identified in four independent families with varying clinical presentations including isolated and combined dystonia ranging from childhood onset to adulthood onset. 3The two patients of our DYS-98 family (IV.3,IV.6) were cousins presenting with early adulthoodonset isolated dystonia.

Genetic findings of the candidate genes in the Turkish dystonia cohort: a reference catalogue for the gene replication studies
Using our variant prioritising strategy, we have prioritised 12 variants in 10 plausible genes with supporting evidence for their possible role in the pathogenesis of dystonia, such as being involved in the CEN modules with the known dystonia genes (table 3).GSEO analysis further implicated those genes in several dystonia-associated pathways and gene ontologies (GOs) (online supplemental material S11).
Through CEN analysis, we identified one (turquoise module; false discovery rate (FDR)-corrected Fisher's exact p=3.62×10 −8 ) out of 13 modules that is statistically enriched for the dystonia-associated genes and involves our candidate genes in the BG region of the brain.Also, the black module (Fisher's exact p=0.04716) in the BG region was classified as a significant module, even though after multiple test correction it was not found to be statistically significant.Our CEN analysis §LIST score ranges between 0 and 1 where higher score indicates more deleterious effect.¶OMIM-based information indicating the involvement of the genes in the diseases.**Metabolic findings suggesting the abnormal levels of associated metabolites.† †ACMG/AMP-based pathogenicity classification.‡ ‡Three families were published previously elsewhere. 28   §LIST score ranges between 0 and 1 where higher score indicates more deleterious effect; >0.

Neurogenetics
also identified one significant module (turquoise module; FDRcorrected Fisher's exact p=0.022383) in the CRBL region.However, no candidate gene was identified in this module.In the putamen region of the UKBEC data, no significant module was identified after FDR correction (online supplemental material S19).

Candidate genes in the CEN modules
Novel compound heterozygous pathogenic variants (NM_000270;c.141del;p.F48LfsTer26) were identified in PNP gene.Structural modelling analysis (online supplemental material S12 revealed that the pathogenic variants have an impact on the ligand binding and enzyme active sites in two affected siblings presenting with paroxysmal movement disorder including ataxia, choreoathetosis and dystonia as well as recurrent severe infection in childhood (family DYS-4).We assessed the effect of the pathogenic variants on the enzyme activity and observed that due to the possible enzyme deficiency, the level of inosine (DYS4-III.1;4.3, DYS4-III.2;1.4 mmol/moK) was elevated in blood.
A novel homozygous pathogenic variant (NM_005914;c.1980G>C;p.R660S) in the MCM4 gene, which was previously associated with immunodeficiency 54 (MIM:609981) (online supplemental material S13), has been found to co-segregate with the disease in a large family (DYS-55) including five unaffected and two affected individuals presenting with adolescence-onset isolated dystonia.Mouse model studies revealed neurological, cardiovascular, embryological and growth impairments in knockout mouse models, as well as lethal impairments in homozygous knockout mouse models, implying that homozygous mutations may cause more severe molecular perturbations.MCM4 is a component of the MCM2-7 complex involved in the initiation of eukaryotic genome replication and has been reported to be implicated in neurogenesis 30 (figure 2C).
A novel homozygous variant (NM_001240;c.1972C>G;p.Q658E) of another compelling gene, CCNT1, was identified in one patient (DYS-80; II.2) presenting with childhood-onset myoclonus dystonia.The CCNT1 gene encodes a subunit of the positive transcription elongation factor B (P-TEFb), or the so-called CDK9/cyclin-T1, which facilitates the elongation of the transcription (figure 2D).

DISCUSSION
This is the first study in the literature that has uncovered the genetic basis or associated clinical characteristics of dystonia families from Turkey using a combination of genetic screening strategies.
Our analyses showed that dystonia is a genetically and clinically highly heterogeneous group of disorders.Even though our study cohort consisted of only recessive forms of the disease from a particular geographical region where recessive forms are common due to the high rate of consanguinity, our analysis revealed a wide spectrum of genetic underpinnings, including known common and uncommon forms of the disease with various modes of inheritance (figure 1).
The highest rates of the variants have been found in the PRKRA and HPCA genes in our dystonia cohort.Findings for HPCA have been discussed before. 28The vast majority of PRKRA pathogenic variant carriers published so far have shown generalised dystonia 31 as our patients do.Based on the ACMG/AMP guidelines, one (p.S235T) of the PRKRA gene variants was not categorised as a pathogenic variant in the group of dystoniaassociated genes due to the incompatible co-segregation in the family (DYS-66), where an asymptomatic family member was also found to be the carrier of the variant.However, the onset of symptoms related to PRKRA pathogenic variants reported so far varies from 1 to 53 years.Therefore, the asymptomatic family member (33 (DYS66; IV.2) year) could also develop symptoms later.
The p.S235T variant has been previously reported to be associated with intellectual disability in infancy. 32Our patient (DYS-66, IV.5) started to suffer from dystonia in the childhood.A neuropsychological examination at the age of 20s revealed mild disinhibition and attentional problems as well as mild verbal memory deficits, but no intellectual disability was reported either for him or for his asymptomatic sister carrying the same variant (p.S235T).Our analysis did not reveal any other plausible variants in the family DYS-66, therefore; the variant was listed in our study.
The p.P222L variant identified together with the p.F68L variant in a compound heterozygous state in the family DYS-86 is the most frequently identified pathogenic variant in PRKRA.Clinical symptoms of this patient progressed quickly, leading to GPi-DBS at age of 20s only 3 years after the onset.
The p.R137H variant in the SGCE gene was not categorised as pathogenic based on the ACMG/AMP guidelines.Due to the clinical compatibility, the variant was also listed in our study.
Our ES approach also identified variants in the autosomal dominant (AD) genes, which include SGCE, KMT2B, GCH1, SLC2A1 and THAP1 (n=5/15; 33.3%).Despite the fact that our cohort consisted of families with an AR mode of inheritance, our study revealed a relatively high proportion of pathogenic or likely pathogenic variants in the AD genes due to (1) the presence of de novo variants, (2) maternal imprinting and (3) the genes' reduced penetrance.

Figure 2
The prioritised pathways associated with dystonia.This figure illustrates the prioritised pathways associated with dystonia based on the findings of our study as well as the findings from previous studies.Those pathways include (A) the endosomal-lysosomal pathway involving AOPEP, VPS16, VPS41 and VPS11 genes, (B) the purine pathway involving IMPDH2, HPRT1 and PNP genes, (C) the cell-cycle pathway including THAP1 and MCM4 genes, (D) the transcriptional regulation pathway including TAF1 and CCNT1 genes.In green rectangles the known dystonia-associated genes, in yellow rectangles candidate genes identified in our study have been shown.HOPS, the homotypic fusion and protein sorting complex; TF, transcription binding factor.

Neurogenetics
In addition, we identified a homozygous truncating variant in AOPEP (C9orf3) gene in a multigeneration family (DYS-98) with two affected individuals presenting with early adulthoodonset isolated dystonia.AOPEP gene has recently been linked to DYT31 causing multifocal or generalised dystonia. 3The pathogenic variant (p.G686EfsTer17) identified in our study is localised in the C-terminal outside of the putative active domains and leads to the early adulthood-onset dystonia, suggesting that the position of the pathogenic variants in the AOPEP gene does not change the impact on the dystonia pathogenesis.Additionally, structural modelling analysis showed that the mutant protein lacked one of the essential ligand binding sites (online supplemental material S14) and had a deficiency in some of the biological functions, such as leukotriene-A4 hydrolase activity, epoxide hydrolase activity and interleukin-1 receptor binding.Our CEN and GSEO analyses implicated AOPEP with some GOs including endocytosis (online supplemental material S11c).Furthermore, PPI analysis showed an interaction (online supplemental material S15b) of AP-O with SUN2, which has been shown as an activator of RAB5 in the endocytic vesicles.Endocytic pathways have been suggested to have a role in differentiation and migration of neuronal cells during development and in adulthood. 33dditionally, a recent study suggested the role of the homotypic fusion and protein sorting complex-mediated endosomallysosomal pathways in the dystonia pathogenesis. 34Therefore, it might be speculated that AOPEP could be implicated in the dystonia pathogenesis through being involved in the endosomallysosomal pathways (figure 2A).
Pathogenic or likely pathogenic variants were found in four families in genes that are not commonly associated with dystonia (PCCB, CACNA1A, ALDH5A1 and PRKN).Among those genes, PCCB and ALDH5A1 have been associated with metabolic disturbances.In line with our findings, we identified downstream metabolic imbalances (table 2), indicating that those pathogenic variants played a role in disease pathogenesis.
In one family (DYS-69), our analysis revealed a de novo pathogenic variant (NM_023035:c.2137G>A:p.A713T) in the CACNA1A gene, which is mainly associated with episodic ataxia and rarely dystonia.It has been proven by AM studies as well as systems-biology-based analysis that dystonia and ataxia share some pathological mechanisms, 35 suggesting the role of convergent disease mechanisms.
Another interesting finding was the identification of homozygous PRKN stop-gain pathogenic variant (NM_004562;c.931C>T;p.Q311X), which was previously reported in a Turkish family presenting with adolescence-onset parkinsonism without dystonia, 36 in a patient with childhoodonset generalised dystonia without parkinsonism.Lower limb dystonia, which sometimes remains an isolated finding for years, is frequently associated with the Parkin-phenotype.However, isolated generalised dystonia is very uncommon in PRKN pathogenic variant carriers, suggesting shared and sometimes convergent mechanisms in different movement disorders, particularly ataxia, dystonia and parkinsonism.
With the advent of NGS techniques, a substantial number of dominantly inherited genes have been associated with dystonia.However, a limited number of recessively inherited genes could be discovered and replicated by independent studies due to the rarity as well as the clinical heterogeneity of AR dystonia.Hitherto, of those genes, only HPCA and recently AOPEP, could be confirmed by independent studies. 4 28 38Therefore, we generated a list of plausible candidate genes with systems-biology-based evidence identified in a clinically well-characterised dystonia cohort to aid in future gene discovery approaches.
Dystonia is defined as a network disorder affecting the BG and the CRBL regions of the brain.Therefore, we generated CENs based on the BG and the CRBL regions (discovery data) from the GTEx database and putamen region (replication data) from the UKBEC database using WGCNA and interrogated the dystonia-associated modules including our candidate genes (n=30).In the CEN analysis, we identified one statistically dystonia-associated module (turquoise) including our candidate genes in the BG region.However, no candidate genes were identified in the significant module (turquoise) of the CRBL region.In the putamen data, no statistically dystonia-associated module was identified.The reason might be that while we could analyse three subregions (caudate, nucleus accumbens and putamen) from the BG region in the GTEx data, we could only analyse putamen subregion in the UKBEC data.Therefore, only the overrepresented pathways, GOs and particular PPIs in the networks for the dystonia-associated modules based on the BG region were determined.Lastly, animal model research-based and literaturebased information with all the evidence supporting the role of those genes in the dystonia pathogenesis were associated with the clinical and genetic findings (table 3).
Based on our strategy, 10 genes (PNP, MCM4, PRDM15, TBC1D32, DZIP3, CCNT1, TBC1D8, ANGEL1, PDF, CEP120) have been prioritised.Of those genes, PNP encodes purine nucleoside phosphorylase and is involved in the purine degradation and salvage pathway (figure 2B).PNP deficiency has been shown to cause a wide range of clinical symptoms, including neurological abnormalities, such as ataxia, tremor, hypertonia and dystonia (in only one case).As has been shown in the case of adenosine deaminase (ADA) deficiency, which is another enzyme involved in the purine degradation-salvage pathway, as well as in the knockout mouse models, purine metabolism could be implicated in neurological disturbances. 39urthermore, our CEN analysis showed that PNP and co-expressed dystonia genes were enriched for immunesystem-associated pathways in particular.Of the associated over-represented pathways, cytokine-cytokine receptor interaction, Jak-STAT signalling pathway and MAPK signalling pathways (online supplemental file S11b) have been shown to be dysregulated in the mutant (THAP1;c.71+9C>A)lymphoblastoid cells. 40In addition, in an independent study, it has been suggested that knockdown of the hypoxanthine guanine phosphoribosyltransferase-a purine biosynthesis gene-led to the constitutive activation of the MAPkinases phosphor-ERK1/2. 41

Neurogenetics
Furthermore, a recently identified new dystonia-associated gene, IMPDH2, supports the role of the purine metabolism in the dystonia pathogenesis. 37 42Based on the recent findings, the association between the purinergic pathways and the neurodevelopmental disorders has been proposed, however, the contribution of those mechanisms to the disease could not be yet elucidated.The possible involvement of the immune systems in the pathogenesis of the purine-metabolism-associated diseases has been reported by different studies, 43 suggesting a presumptive indication of a link with dystonia pathogenesis.
The genes in the CEN modules identified in our study were enriched also for the transcription processes, ubiquitinationrelated mechanisms and synaptic transmission, which are the molecular mechanisms disturbed in dystonia 44 45 (online supplemental material S11a).
Of those genes, MCM4 has been shown to be involved in DNA replication complex with a particular role in neurogenesis, and furthermore, CEN and PPI analysis showed that MCM4 and Cdc7 (online supplemental material S15a) might act as a complex in cell-cycle regulation, suggesting a possible role of the MCM4 gene in the cell-cycle-related dystonia pathogenesis (figure 2C).Another compelling gene, CCNT1 is involved in the elongation of the transcription and reported to be an interactor 46 of TAF1 (dystonia-parkinsonism; MIM 314250) (figure 2D).
We also identified variants in the genes (n=20) without CENbased evidence due to their low or different expression pattern in different brain regions and their involvement in different pathogenic mechanisms, where the known dystonia-associated genes are not involved (online supplemental material S17).Some of those genes have been reported to be involved in neurological disturbances in AM studies.Therefore, we did not rule out those genes as possible genetic causes in the affected families in our cohort.
Our study failed to retrieve variants in seven families (17%) (for the detailed clinical information, see online supplemental material S4), suggesting either the limitations of the ES-based screening approach or some inconsistency in the clinical information obtained from the families.
In conclusion, in this study, we have uncovered the genetic basis in dystonia families with well-defined clinical characteristics from Turkey.Using comprehensive genetic screening strategies, we genetically and clinically characterised the families with pathogenic variants in the known dystonia-associated genes, and provided a list of candidate genes with CEN-based, animal research-based, PPI-based or literature-based evidence for the pathogenicity of dystonia.Our study revealed several insights into the pathogenesis of dystonia, including: (1) the substantial difference in the genetic composition of dystonia in different populations; (2) the importance of broad genetic screening approaches in diagnostic settings rather than focusing only on known dystonia genes; (3) the considerable involvement of convergent disease mechanisms in the pathogenesis of dystonia, ataxia, and Parkinson's disease and (4) the significant contribution of novel and known pathways, such as immune system, transcription, neurodevelopmental, endosomal-lysosomal and metabolic pathways.

Figure 1 2 *
Figure 1 The classification and the characterisation of the 42 Turkish dystonia families based on the genetic findings.The chart indicates (A) the classification of the families based on the genetic findings, (B) the clinical distribution of the families in the diagnostic group and (C) the variant distribution in the diagnostic group.

NM_152730*
GnomAD-based MAF and allele counts.†CADD score of >15 indicates deleteriousness for the variant.‡M-CAP is a pathogenicity classifier for rare missense variants and S-CAP for the splicing variants.

Table 1
Summary of the clinical characteristics of the dystonia cohort

Table 3
The characteristics of the prioritised variants in the candidate genes with CEN-based, PPI-based and AM-based evidence 6 is considered as deleterious effect.¶OMIM and literature findings were indicated if the mode of inheritance of the associated disease is compatible.