The study subjects were members of three generations of a Han Chinese ataxia family pedigree in Henan province. A detailed medical history and physical examination record of the proband (II₅) and some family members (II₇, II₉, II₁₁, III_13, and III₁₅) were evaluated by two experienced neurologists. Peripheral blood specimens were obtained from the proband (II₅) and some members (II_7, II₉, II₁₁, II₁₃ II₁₅, III₄, and III₁₀) for genetic testing. This study was approved by the local ethics committee, and all patients signed an informed consent form.

Clinical physical examinations and evaluations were performed in the proband (II₅) and other members (II₉, II₁₁, and II₁₅), including the Scale for Assessment and Rating of Ataxia (SARA),Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), Hamilton Anxiety Scale (HAMA), Hamilton Depression Scale (HAMD) and Pittsburgh Sleep Quality Index (PSQI). Electrophysiological examinations conducted include; nerve conduction velocity (NCV), electromyography (EMG), motor evoked potential (MEP), somatosensory evoked potential (SSEP), visual evoked potential (VEP), brainstem auditory response (BAEP) and pure-tone audiometry (PTA). However, II₇ and II₁₃ received only clinical history questioning, physical examination, and scale assessment.

The proband (II₅) and three family members (II₉, II₁₁, and II₁₅) underwent a detailed MRI examination. Among them, II₇ underwent MRI flat-scan examination before one month, and the image acquisition was performed using Siemens Magnetom Prisma 3.0T MRI scanner in the following sequence: T1WI, T2WI, FLAIR, DWI, 3D-T1, and DTI images.

3D-T1 images were segmented in the MNI space using a Brain Label. (Ye et al., 2018) The whole brain was segmented into different anatomical structures using a Brain Label with 283 optimal regions. Brain atlases were transformed into local space by aligning 3D-T1 images using a symmetric differential isomorphic image alignment algorithm built into ANTs. We quantitatively analyzed the segmented brain regions of each patient to obtain the volume of each brain region. Brain volume data were then compared with those of healthy individuals of the same sex and age according to the Brain Label database.

DTI image preprocessing was performed using the PANDA to generate FA DEC and AD maps. (Wasserthal et al., 2019) The automated fiber quantification analysis of white matter fibers was then carried out. The data obtained for FA values are expressed as mean ± standard deviation. SPSS 22.0 statistical software was applied to statistically analyze the measurements at different sites bilaterally and paired t-test was used for the bilateral comparisons. p < 0.05 was considered to be statistically significant.

Due to the initial suspicion of SCA, the RP-PCR and capillary electrophoresis were performed to identify SCA subtypes of the proband (II₅) and some members (II_7, II₉, II₁₁, II₁₃, II₁₅, III₄, and III₁₀), including SCA1, 2, 3, 6, 7, 8, 10, 12, 17, 36, and DRPLA. RP-PCR relies on repeat primers with amplified alleles that anneal to produce the results for PCR products separated by capillary electrophoresis is “ladder”. For DNA fragment analysis, RP-PCR products were analyzed using the ABI-Prism 3730XL Genetic Analyzer, and the data were analyzed using GeneMarker software.

The proband (II₅) and two other members (II₁₁ and II₁₅) were subjected the exome sequencing through Illumina HiSeq platform. Single nucleotide variants (SNV)and insertions and deletions (InDels) were analyzed by Genome Analysis Tool Kit (GATK) software. Single nucleotide variants (SNV) and insertions and deletions (InDels) were analyzed by GATK software. Genome Analysis Tool Kit Variants with minor allele frequencies >0.5% were filtered out by the databases, including the gnomAD, ExAC and 1000 genome databases. Functional prediction of candidate variants was performed using SIFT, Polyphen-2 and Mutation Taster software. All variants were interpreted according to ACMG recommendations based on the ClinVar, OMIM, and HGMD databases. Sanger sequencing was used to validate the genetic variants detected by exome sequencing.

We selected the three samples (II₅, II₇, and II₁₁), which presented different clinical symptoms, and were sequenced by long-read genome sequencing. PacBio CLR sequencing has an error rate of 11%–15%. The sequencing errors are random and can be corrected by increasing the coverage of sequencing. To obtain more full-length subreads to get the exact repeat region length and interruptions landscape, we set the coverage sequencing to ×100.1. Genomic DNA Sample Preparation: Samples were collected, and high molecular weight genomic DNA was prepared by the CTAB method and followed by purification with QIAGEN^® Genomic kit (Cat#13343, QIAGEN) for regular sequencing, according to the standard operating procedure provided by the manufacturer. The DNA degradation and contamination of the extracted DNA was monitored on 1% agarose gels. DNA purity was then detected using NanoDrop™ One UVVis spectrophotometer (Thermo Fisher Scientific, USA), of which OD260/280 ranging from 1.8 to 2.0 and OD 260/230 is between 2.0 and 2.2. At last, DNA concentration was further measured by Qubit^® 4.0 Fluorometer (Invitrogen, USA). 2. Library preparation and sequencing: The SMRTbell Continuous Long Read (CLR) library was constructed for sequencing according to PacBio’s standard protocol (Pacific Biosciences, CA, USA) using either 10 kb or 20 kb preparation solutions. The main steps for library preparation are: 1) gDNA shearing, 2) DNA damage repair, end repair and A-tailing, 3) ligation with hairpin adapters from the SMRTbell Express Template Prep Kit 2.1 (Pacific Biosciences), 4) size selection, and 5) binding to polymerase. Briefly, a total amount of 5 μg DNA per sample was used for the DNA library preparations. The genomic DNA sample was sheared by g-TUBEs (Covaris, USA) according to the expected size of the fragments for the library. Single-strand overhangs were then removed, and DNA fragments were damage repaired, end repaired and A-tailing. Then the fragments ligated with the hairpin adaptor for PacBio sequencing. Target fragments were screened by the BluePippin (Sage Science, USA). The SMRTbell library was then purified by AMPure PB beads, and Agilent 2,100 Bioanalyzer (Agilent technologies, USA) was used to detect the size of library fragments. Sequencing was performed on a PacBio Sequel II instrument with Sequencing Primer V4 and Sequel II Binding Kit 2.1 in Haorui Genomics. After sequencing, we take the effective filtering strategies to improve the sequencing data quality. The filtered reads were assembled into the individual genome using NextDenove software. Then, we mainly analyzed the GGCCTG repeat region in the intron 1 region of the NOP56 gene. To better observe the structural variation, the genome GRCh38 was artificially modified by inserting d (GGCCTG)₁₀₀₀ before the original d (GGCCTG)₄ to construct a fake GRCh38 sequence. The subreads were then compared to GRCh38 and fake GRCh38. The GGCCTG repeat number in each sample were counted, and the number of subreads containing GGCCTG repeat number greater than 650 and the full-length subreads of GGCCTG repeat number and length were also analyzed. Based on other about SCA studies, except for core motif GGCCTG, the pathogenic regions exist interruption motifs in similar diseases, such as SCA10, so it is may likely that not all regions actually obtained are composed of GGCCTG tandem repeats. (Matsuura et al., 2006; McFarland et al., 2013) In order to visualize the nucleotide sequence, the schematic representation of the motifs was produced based on the Practical Extraction and Report Language. Statistical analysis of the tandem repeat motifs in the d (GGCCTG) n region by the schematic representation of all full-length subreads were performed to determine the distribution regularity of the motifs.

The family pedigree of ataxia is shown in (Figure 1). 6 patients and 2 presymptomatic individuals were identified in the pedigree, and the detailed data of each affected member are shown in Table 1. The proband (II₅) was a 62-year-old female. At age 40, this patient was observed to have affective disorders manifested by low motivation and irritability, along with severe insomnia and long-term use of valium. At age 47, the patient began experiencing unstable walking. At age 49, her speech became increasingly incoherent, and she occasionally choked on water, along with self-perceived hearing loss and inability to clearly hear what others said. At age 58, she developed blurred vision, occasional dizziness, and memory loss. The patient was 62 years old at the time of evaluation and showed cognitive impairment (MMSE: 22/30); (MoCA: 12/30), with mild depression, anxiety and sleep disturbances.

To summaries the clinical characteristics of the family, four female and two male patients were evaluated. The main clinical symptoms were ataxia (5/6, 83.3%), insomnia (4/6, 66.7%), dysarthria (3/6, 50%), affective disorders (3/6, 50%), blurred vision (3/6, 50%), positive pathological signs (3/6, 50%), hearing loss (2/6, 33.3%), tongue muscle atrophy and muscle bundle tremor (1/6, 16.7%). (Table 1)

In electrophysiological examinations, two patients (II₅ and II₁₁) had central and peripheral damage to the auditory pathway as revealed by BAEP, which showed the disappearance of I–III waves and normal V waves or prolonged I–III or III–V interpeak latency. Two patients (II₅ and II₁₁) had hearing loss, as shown in the PTA test, mainly in the high-frequency hearing threshold. One patient (II₁₁) had peripheral nerve damage Two patients (II₅ and II₁₁) had positive pyramidal tract sign by MEP (Supplementary Figure S1) However, all serum examination results were within the normal range.

Two experienced imaging physicians interpreted and processed the images. They suggested that 3 patients (II_5, II₇ and II₁₁) had visual microcephaly, and II₉ was assessed as normal (Figure 2D). 3D-T1 analysis showed that the patient had reduced cerebellar volume, less cerebellar white matter volume, and reduced pons volume than the controls (Figures 2A, B, Supplementary Table S1). DTI analysis also showed that the patient had reduced FA values in the superior and inferior peduncles of the left cerebellum compared to the controls (corrected p < 0.05) (Figure 2C; Supplementary Table S2).

Because of the clinical characteristics of SCA in this family, genetic testing was used to determine the SCA subtype. The NOP56 capillary electrophoresis map of the proband (II₅) showed a single peak (Figure 1B), and a characteristic ladder pattern with a 6-bp periodicity on the electropherogram was identified by RP-PCR (Figure 1C). The results for the other affected members displayed the same features. We also found the abnormal CAG expansion repeat of the SCA17-associated TBP gene in this pedigree (Figure 1D), following the verification by Sanger sequencing that the CAG/CAA repeat number was 43. It is reported that most patients carry intermediate TBP _41-49 alleles that show incomplete penetrance. Also, the parkinsonism, dystonia, seizure and chorea were typical features occurred in SCA17 can’t observed in this pedigree. A study showed that the ataxia-related phenotype occurs when TBP is in the incomplete penetrance genotype and also has STUB1 variants. (Magri et al., 2022) However, we did not detect STUB1 variants in the affected family members. Therefore, we considered that the aberrant CAG repeat expansion of the TBP in this family did not contribute to the clinical phenotype in this family. The results of RP-PCR are shown in (Supplementary Table S2). WES revealed no mutations in genes associated with ataxia in this pedigree.

The results of the SMRT sequencing data for the three samples are shown in (Supplementary Table S4). To avoid errors in SMRT sequencing technology and increase the reliability of sequencing sequences, the sequencing coverage was set to ×100, and the raw data reached more than 300 Gb. We focused on the structural variation analysis of the assembled genome for intron 1 of the NOP56 gene and its sequences on both sides. The filtered subreads were then aligned to GRCh38 and fake GRCh38 reference genomes. Based on the comparison, subreads that were completely or partially across d (GGCCTG) n were extracted. The subreads were sorted into full_dGGCCTGn, 5p_GGCCTGn, 3p_GGCCTGn and full_dGGCCTGr, and the statistics of each subread were recorded (Table 2; Figure 3A). The GGCCTG motif was found in all three samples, and subreads with more than 650 repeats were counted in each sample (Table 2). The three samples had more than 650 subreads, indicating that all three carried pathogenic repeat variants and could be clearly diagnosed as SCA36.

We analyzed the genotype and phenotype in the patients. The full_dGGCCTGn subreads of three patients (II₅, II₇, and II₁₁) were 15,27, and 24, respectively, and the length of each subread was different. The average (repeat region) length of II₅ was 5,988 bp and the average repeat (unit) number was 598, the average length of II₇ was 7,663 bp and the average number of repeats was 787; while the average length of II₁₁ was 8,849 bp and the average number of repetitions was 868 (Figures 3B, C). The proportion of subreads with more than 650 repeat number was 11.84%, 14.20%, and 20.83% for II₅, II₇, and II₁₁, respectively. SARA score was used to assess the severity of ataxia symptoms in patients. The scores of the three patients (II₅, II₇, and II₁₁) were 10, 8.5, and 8.5, respectively. However, we considered that the disease duration of the three patients (II₅, II₇, and II₁₁) is approximately 16, 10 and 7 years. From the perspective of disease progression, II₁₁ had the fastest progression, followed by II₇, and II₅ the slowest. It is reported that the repeat size is associated with the clinical features in the disease with repeat expansion like Huntington’s disease. (Trottier et al., 1994) Combining genotype-phenotype analysis, we speculated that the average length, the average number of repeats of full_dGGCCTGn subreads, and the proportion of subreads with more than 650 repeats were associated with the age of onset and disease progression.

The motif structure of long-read sequencing study in NOP56 repeat expansion will be important to determine heterogeneity and whether the repeats are interrupted by non-GGCCTG content. Therefore, we further analyzed the composition and distribution of motifs within the repeat expansion regions of the three samples. The results showed that both the number and proportion of motifs were dominated by GGCCTG, and the rest were GGCTG, GGCCCTG, GGCCG, and GGCCTTG (Figure 4C). It was also found that the motifs mostly consisted of 5-nt, 6-nt, and 7-nt. It suggested that SCA36 patients had interruptions in the repeat expansion region, and the motifs varied around the core motif of GGCCTG. However, the TOP10 motifs, i.e., 5 nt, 6 nt, and 7 nt motifs, were not significantly different among the three samples (Figures 4B–D).

We performed IGV visualization analysis of each full-length subread of the three patients and produced correlation schematics based on the Practical Extraction and Report Language, which showed no obvious regularity in the interruptions. We selected the full-length schematics with the highest GGCCTG repeat number from each of the three patients for analysis (Figures 5A–C) The various motifs showed a free insertion pattern with no obvious regularity; therefore, we think that the position of the interruptions may not be clearly associated with the phenotype of the patients in this pedigree.