Abstract
Our study provides new insights into the potential causative factors of WS and an alternative way to explore clinically undiagnosed cases, which will promote clinical diagnosis and genetic counseling. However, the two potential disease-causing genes (KIT, CHD7) and 32 potential pathogenic variants (PAX3: 20, MITF: 7, SOX10: 5) predicted by multi-data integration in this study are all computational predictions and need to be further verified through experiments in follow-up research.
Introduction
Waardenburg syndrome (WS) is a rare genetic disorder characterized by sensorineural hearing loss and pigmentary abnormalities, such as piercing blue eyes or a white forelock. While seven causative genes have been identified, approximately 38.9% of patients remain molecularly unexplained after standard genetic testing. Furthermore, WS is highly heterogeneous; patients with different genetic mutations show vastly different symptoms, and existing genotype-phenotype studies have often been limited by small sample sizes or regional biases. There is a critical need to identify missing genetic causes and better understand how specific mutations translate into clinical symptoms.
Methodology
The researchers employed a multi-data integration approach to identify new candidate genes and variants. The pipeline combined protein-protein interaction (PPI) networks, phenotype similarity analysis, and single-cell RNA sequencing to hunt for new causative factors.
The workflow involved several key steps:
- Multi-Data Integration: Known WS genes were mapped into three major PPI databases (BioGRID, InWeb_InBioMap, and HitPredict). Phenotype similarity scores were calculated using the Human Phenotype Ontology (HPO). Genes that both interacted with known WS proteins and shared similar clinical symptoms were flagged.
- Biological Validation: Candidates were cross-referenced with the MORL scRNA-Seq database to check expression in cochlear hair cells and the Mouse Genome Informatics (MGI) database to identify hearing loss phenotypes in knockout models.
- Variant Prediction: Using ANNOVAR and five functional prediction algorithms (SIFT, PolyPhen2, MutationTaster, CADD, and GERP++), thousands of possible mutations in PAX3, MITF, and SOX10 were screened to identify 32 highly likely pathogenic variants.
- Clinical Correlation: A dataset of 443 WS cases from 84 published studies was curated to perform a large-scale statistical analysis of genotype-phenotype correlations.
Key Insights and Results
The study identified two new potential pathogenic genes (KIT and CHD7) and 32 potential pathogenic variants (20 in PAX3, 7 in MITF, and 5 in SOX10).
Genotype-Phenotype Mapping
The large-scale analysis revealed distinct clinical profiles associated with specific genes:
- PAX3: Strongly associated with white forelock and telecanthus (widely spaced inner corners of the eyes). A wide nasal bridge was found to be exclusive to PAX3 variants.
- MITF: Associated with a higher frequency of skin freckles and premature graying.
- SOX10: Dominant in cases involving constipation and aganglionic megacolon (Hirschsprung’s disease). SOX10 variants also carried a significantly higher risk of neurological and auditory-specific diseases, correlating with its higher expression in brain and ear tissues.
Mutation Distribution
The study confirmed that 95.98% of causative variants occur in exon regions, supporting the continued use of whole-exome sequencing (WES) as a primary diagnostic tool. Statistical analysis showed no significant difference in phenotype prevalence between male and female patients.
Conclusion and Future Directions
This research highlights that disease-causing genes rarely act in isolation. By examining the "neighbors" of known WS genes in protein networks, the study successfully identified high-priority candidates. However, these findings are computational predictions. While supported by expression data and mouse models, the identified genes and variants require "wet lab" validation, such as CRISPR-edited animal models, to definitively prove causality.
Since nearly 40% of cases remain unexplained, future research should also explore non-coding regions. Utilizing RNA-seq to detect deep intronic mutations, splicing defects, or fusion transcripts may uncover the remaining "missing" genetic causes of Waardenburg syndrome.
Summary updated