UMMAN: Unsupervised Multi-graph Merge Adversarial Network for Disease Prediction Based on Intestinal Flora

TL;DR

This work tackles disease prediction from gut microbiome data by modeling multiplex host-microbe relationships in an unsupervised setting. It introduces UMMAN, which constructs a Original-Graph and a Shuffled-Graph across multiple relation-types, learns node embeddings with Graph Neural Networks, and aggregates them through a Node Feature Global Integration descriptor and an attention-based fusion, guided by a joint loss that combines $L_{adv}$ and $L_{h-attn}$. The key contributions are the first integration of GNNs into unsupervised gut microbiome disease prediction, the use of a multi-graph with a disruption-based adversarial signal, and the NFGI-based global graph descriptor, validated on five OTU datasets where UMMAN achieves state-of-the-art and stable performance. This approach advances disease diagnosis potential from gut microbiota by capturing complex multiplex interactions across hosts, with practical impact on bioinformatic pipelines for microbiome-based prognosis.

Abstract

The abundance of intestinal flora is closely related to human diseases, but diseases are not caused by a single gut microbe. Instead, they result from the complex interplay of numerous microbial entities. This intricate and implicit connection among gut microbes poses a significant challenge for disease prediction using abundance information from OTU data. Recently, several methods have shown potential in predicting corresponding diseases. However, these methods fail to learn the inner association among gut microbes from different hosts, leading to unsatisfactory performance. In this paper, we present a novel architecture, Unsupervised Multi-graph Merge Adversarial Network (UMMAN). UMMAN can obtain the embeddings of nodes in the Multi-Graph in an unsupervised scenario, so that it helps learn the multiplex association. Our method is the first to combine Graph Neural Network with the task of intestinal flora disease prediction. We employ complex relation-types to construct the Original-Graph and disrupt the relationships among nodes to generate corresponding Shuffled-Graph. We introduce the Node Feature Global Integration (NFGI) module to represent the global features of the graph. Furthermore, we design a joint loss comprising adversarial loss and hybrid attention loss to ensure that the real graph embedding aligns closely with the Original-Graph and diverges from the Shuffled-Graph. Comprehensive experiments on five classical OTU gut microbiome datasets demonstrate the effectiveness and stability of our method. (We will release our code soon.)

Figures (4)

• Figure 1: The architecture of the UMMAN we propose. The subgraph on the left is the core architecture of the entire network. The subgraph on the right is the specific process of the NFGI module, including Node-level stage and Graph-level stage.
• Figure 2: Abundance correlations of the top 10 most abundant gut microbes in the Cirrhosis dataset.
• Figure 3: Intuitive comparison of our method with previous work on the five OTU datasets.
• Figure 4: The comparison on four variants of our method.