X2Graph for Cancer Subtyping Prediction on Biological Tabular Data

2025-05-29

AI Summary

Modeling few-shot biomedical tabular data—such as cancer gene expression profiles—is challenging due to limited samples and the absence of structural priors. Method: We propose a knowledge-enhanced graph neural network framework that explicitly encodes inter-column biological priors (e.g., gene–gene interaction networks) into sample-level structured graphs, enabling interpretable tabular-to-graph transformation; it then fuses local features with external knowledge via graph message passing. Contribution/Results: This approach bridges tabular learning and graph representation learning, specifically designed for low-data biomedical tasks. Evaluated on three cancer subtype prediction datasets, it significantly outperforms strong baselines—including XGBoost and TabNet—with AUC improvements of 5.2–8.7 percentage points in few-shot settings. Results demonstrate that incorporating structured biological priors substantially enhances the generalization capability of deep tabular models.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Despite the transformative impact of deep learning on text, audio, and image datasets, its dominance in tabular data, especially in the medical domain where data are often scarce, remains less clear. In this paper, we propose X2Graph, a novel deep learning method that achieves strong performance on small biological tabular datasets. X2Graph leverages external knowledge about the relationships between table columns, such as gene interactions, to convert each sample into a graph structure. This transformation enables the application of standard message passing algorithms for graph modeling. Our X2Graph method demonstrates superior performance compared to existing tree-based and deep learning methods across three cancer subtyping datasets.

Problem

• Predicts cancer subtypes using biological tabular data
• Improves deep learning for small medical datasets
• Converts tabular data to graphs using gene interactions

Innovation

• Converts tabular data into graph structures
• Leverages external knowledge for column relationships
• Applies message passing for graph modeling