Interpretable Graph Kolmogorov-Arnold Networks for Multi-Cancer Classification and Biomarker Identification using...

Area of Science:

• Computational biology
• Bioinformatics
• Machine learning in oncology

Background:

• Integrating heterogeneous multi-omics data for precision cancer diagnostics is a significant computational challenge.
• Existing analytical models struggle with the complexity and dimensionality of multi-omics datasets.

Purpose of the Study:

• To introduce Multi-Omics Graph Kolmogorov-Arnold Network (MOGKAN), a novel deep learning framework for cancer classification.
• To leverage messenger RNA, micro RNA, DNA methylation, and Protein-Protein Interaction networks for improved diagnostic accuracy.
• To enhance the interpretability of multi-omics data analysis in cancer.

Main Methods:

• Utilized differential gene expression analysis with DESeq2, LIMMA, and LASSO regression for data dimensionality reduction.
• Developed a deep learning architecture based on the Kolmogorov-Arnold theorem principle with trainable univariate functions.
• Integrated messenger RNA, micro RNA, DNA methylation, and Protein-Protein Interaction networks for classification across 31 cancer types.

Main Results:

• Achieved a high cancer classification accuracy of 96.28 percent.
• Demonstrated low experimental variability compared to other deep learning models.
• Identified and validated cancer-related biomarkers through Gene Ontology and KEGG enrichment analysis.

Conclusions:

• MOGKAN effectively integrates multi-omics data with graph-based deep learning for robust cancer classification.
• The framework offers enhanced interpretability, facilitating the translation of complex data into clinically actionable insights.
• MOGKAN shows potential for advancing precision cancer diagnostics.