DistSNE: Distributed computing and online visualization of DNA methylation-based central nervous system tumor...

Area of Science:

• Neuro-oncology
• Computational Biology
• Genomics

Background:

• Current central nervous system (CNS) tumor classification relies on centralized DNA methylation profiling.
• Centralized approaches raise privacy concerns regarding user data.
• Existing methods require data aggregation, limiting scalability and collaboration.

Purpose of the Study:

• To develop a distributed computing approach for CNS tumor classification.
• To achieve performance comparable to centralized methods while ensuring data privacy.
• To provide a user-friendly platform for collaborative methylome analysis.

Main Methods:

• Utilized a distributed computing framework (DataSHIELD) for federated analysis.
• Employed t-distributed neighborhood embedding (t-SNE) for dimensionality reduction and visualization (DistSNE).
• Developed an intuitive web interface for local data management and classification.

Main Results:

• DistSNE achieved comparable performance to centralized CNS tumor classification methods.
• The distributed approach successfully safeguarded user data privacy.
• Increased sample size through distributed data improved cluster analysis and predictive power.

Conclusions:

• DistSNE offers a simple, fast, and privacy-preserving method for CNS tumor classification using large-scale methylation data.
• The platform facilitates collaboration among multiple institutions, enhancing neuro-oncology research.
• This approach holds significant potential for advancing precision medicine in brain tumor classification.