MOGONET integrates multi-omics data using graph convolutional networks allowing patient classification and biomarker identification

Tongxin Wang,Kun Huang,Jie Zhang,Wei Shao,Zhi Huang,Haixu Tang,Zhengming Ding

doi:10.1038/s41467-021-23774-w

Abstract

To fully utilize the advances in omics technologies and achieve a more comprehensive understanding of human diseases, novel computational methods are required for integrative analysis of multiple types of omics data. Here, we present a novel multi-omics integrative method named Multi-Omics Graph cOnvolutional NETworks (MOGONET) for biomedical classification. MOGONET jointly explores omics-specific learning and cross-omics correlation learning for effective multi-omics data classification. We demonstrate that MOGONET outperforms other state-of-the-art supervised multi-omics integrative analysis approaches from different biomedical classification applications using mRNA expression data, DNA methylation data, and microRNA expression data. Furthermore, MOGONET can identify important biomarkers from different omics data types related to the investigated biomedical problems.

Highlights

To fully utilize the advances in omics technologies and achieve a more comprehensive understanding of human diseases, novel computational methods are required for integrative analysis of multiple types of omics data
We demonstrated the capabilities and versatility of Multi-Omics Graph cOnvolutional NETworks (MOGONET) through a wide range of biomedical classification applications, including Alzheimer’s disease patient classification, tumor grade classification in low-grade glioma (LGG), kidney cancer type classification, and breast invasive carcinoma subtype classification
We showed the necessity of integrating multiple omics data types as well as the importance of combining both graph convolutional networks (GCN) and View Correlation Discovery Network (VCDN) for multi-omics data classification through comprehensive ablation studies

Summary

Introduction

To fully utilize the advances in omics technologies and achieve a more comprehensive understanding of human diseases, novel computational methods are required for integrative analysis of multiple types of omics data. Huang et al.[2] integrated the features of mRNA expression and miRNA expression data with additional clinical information at hidden layers for better prognosis prediction in breast cancer These existing methods are based on fully connected networks, which did not exploit the correlations between samples effectively through similarity networks. It is crucial to utilize the correlations across different classes and different omics data types to further boost the learning performance To this end, we introduce MOGONET, a multi-omics data analysis framework for classification tasks in biomedical applications. To the best of our knowledge, MOGONET is the first supervised multi-omics integrative method that utilizes GCNs for omics data learning to perform effective class prediction on new samples. We demonstrated that MOGONET can identify important omics signatures and biomarkers related to the investigated biomedical problems

Objectives

Methods

Results

Conclusion