Abstract:Accurate cancer subtype prediction is crucial for personalized medicine. Integrating multi-omics data represents a viable approach to comprehending the intricate pathophysiology of complex diseases like cancer. Conventional machine learning techniques are not ideal for analyzing the complex interrelationships among different categories of omics data. Numerous models have been suggested using graph-based learning to uncover veiled representations and network formations unique to distinct types of omics data to heighten predictions regarding cancers and characterize patients' profiles, amongst other applications aimed at improving disease management in medical research. The existing graph-based state-of-the-art multi-omics integration approaches for cancer subtype prediction, MOGONET, and SUPREME, use a graph convolutional network (GCN), which fails to consider the level of importance of neighboring nodes on a particular node. To address this gap, we hypothesize that paying attention to each neighbor or providing appropriate weights to neighbors based on their importance might improve the cancer subtype prediction. The natural choice to determine the importance of each neighbor of a node in a graph is to explore the graph attention network (GAT). Here, we propose MOGAT, a novel multi-omics integration approach, leveraging GAT models that incorporate graph-based learning with an attention mechanism. MOGAT utilizes a multi-head attention mechanism to extract appropriate information for a specific sample by assigning unique attention coefficients to neighboring samples. Based on our knowledge, our group is the first to explore GAT in multi-omics integration for cancer subtype prediction. To evaluate the performance of MOGAT in predicting cancer subtypes, we explored two sets of breast cancer data from TCGA and METABRIC. Our proposed approach, MOGAT, outperforms MOGONET by 32% to 46% and SUPREME by 2% to 16% in cancer subtype prediction in different scenarios, supporting our hypothesis. Our results also showed that GAT embeddings provide a better prognosis in differentiating the high-risk group from the low-risk group than raw features.

Graph Neural Networks with Multiple Prior Knowledge for Multi-Omics Data Analysis

An explainable graph neural network approach for effectively integrating multi-omics with prior knowledge to identify biomarkers from interacting biological domains

Prior knowledge-guided multilevel graph neural network for tumor risk prediction and interpretation via multi-omics data integration

A Multimodal Graph Neural Network Framework of Cancer Molecular Subtype Classification

Comparative Analysis of Multi-Omics Integration Using Advanced Graph Neural Networks for Cancer Classification

AutoGGN: A Gene Graph Network AutoML Tool for Multi-Omics Research

A multimodal graph neural network framework for cancer molecular subtype classification

Graph neural networks for multi-view learning: a taxonomic review

MetaGXplore: Integrating Multi-Omics Data with Graph Convolutional Networks for Pan-cancer Patient Metastasis Identification

Knowledge-driven multi-graph convolutional network for brain network analysis and potential biomarker discovery

MoGCN: A Multi-Omics Integration Method Based on Graph Convolutional Network for Cancer Subtype Analysis

Graph Neural Networks and Their Current Applications in Bioinformatics

A Multi-Omics Integration Framework Using Graph Attention Networks for Cancer Subtype Prediction

Integration of multi-omics data using adaptive graph learning and attention mechanism for patient classification and biomarker identification

A data-centric framework of improving graph neural networks for knowledge graph embedding

Knowledge-guided Learning Methods for Integrative Analysis of Multi-omics Data

PT-KGNN: A framework for pre-training biomedical knowledge graphs with graph neural networks

Molecular Subtyping of Cancer Based on Robust Graph Neural Network and Multi-Omics Data Integration

Multi-omics Graph Knowledge Representation for Pneumonia Prognostic Prediction

Multi-omics data integration by generative adversarial network

Integration of Multi-Omics Data for Gene Regulatory Network Inference and Application to Breast Cancer