Comprehensive Characterization of Clinical Phenotypes and CTG Repeat-Associated Manifestations in Spinocerebellar Ataxia Type 8.

Authors

Deli Yang, Chao Tang, Yu Zhan, Lang Yang, Xiaoyang Lei, Xiaoxue Peng, Dian He

Published in

Cerebellum (London, England). Volume 25. Issue 4. Jul 22, 2026. Epub Jul 22, 2026.

Abstract

Spinocerebellar ataxia type 8 (SCA8) is traditionally characterized as a cerebellar syndrome, yet emerging evidence suggests significant phenotypic heterogeneity. The complex non-linear relationship between genetic burden and clinical severity remains understudied. We analyzed 172 SCA8 patients using advanced data mining techniques. Complete clinical demographic data were available for the cohort, with precise genetic repeat numbers available for 171 individuals. K-means clustering identified patient subtypes, while Gaussian Graphical Models (GGM) visualized symptom topology. We developed exploratory machine learning models (Decision Tree/Random Forest) and employed Generalized Additive Models (GAM) to characterize non-linear genotype-phenotype dynamics. Clustering identified three distinct phenotypes: Severe Multi-system (12%), Mild/Atypical (38%), and Classical Cerebellar (50%). Network analysis revealed ataxia as a central hub, while ocular motor disorders acted as bridges between symptom clusters. A machine learning-derived decision tree stratified patients with promising exploratory performance (AUC: 0.82-0.89) using only three clinical checkpoints (pyramidal signs, speech disorders, and hyperreflexia), showing predictive patterns distinct from genetic repeat length. Additionally, GAM analysis identified a potential non-linear inflection point at ~ 100 CTG repeats. This hypothesis-generating observation provides a preliminary mathematical context for the paradoxical "low-repeat/high-severity" phenotype in the Severe cluster. SCA8 is a multi-dimensional spectrum disorder rather than a uniform cerebellar disease. Our exploratory diagnostic modeling provides a computational framework for understanding symptom predictors, while non-linear modeling provides novel insights into pathogenic heterogeneity, offering an exploratory foundation for future personalized management and clinical trial design.

PMID:
42484951
Bibliographic data and abstract were imported from PubMed on 22 Jul 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.