Applications of node-based resilience graph theoretic framework to clustering autism spectrum disorders phenotypes

Area of Science:

• Graph theory
• Machine learning
• Biomedical data analysis

Background:

• Network data and graph clustering are crucial for understanding complex datasets.
• Autism Spectrum Disorder (ASD) presents significant phenotypic heterogeneity.
• Identifying subgroups within ASD is essential for targeted research and understanding.

Purpose of the Study:

• To apply a novel graph theoretic clustering method, Node-Based Resilience clustering (NBR-Clust), to Autism Spectrum Disorder (ASD) data.
• To identify meaningful subgroups within the ASD population.
• To discover relevant biomarkers for understanding ASD heterogeneity.

Main Methods:

• Utilized Node-Based Resilience clustering (NBR-Clust), a novel graph theoretic approach.
• Constructed graphs to represent ASD phenotype data from the Simons Simplex Collection (SSC) dataset.
• Employed graph quality measures, internal cluster validation, and feature extraction for analysis.

Main Results:

• The NBR-Clust method successfully identified subgroups within the ASD data.
• Analysis revealed potential biomarkers that characterize these subgroups.
• The optimal clustering configuration predominantly resulted in 5 distinct subgroups.
• Demonstrated the utility of resilience measure clustering for biomedical datasets.

Conclusions:

• Node-Based Resilience clustering is a promising method for analyzing heterogeneous biomedical data, specifically for Autism Spectrum Disorder.
• The identified subgroups and biomarkers offer new insights into ASD phenotypic heterogeneity.
• This approach can advance future research and understanding of ASD.