Gut Microbiota Analysis Reveals Microbial Signature for Multi-Autoimmune Diseases Based on Machine Learning Model

Tianfeng An¹Shuya Zhang²Jinjin Li³Hui Wang¹Li Chen¹Yiran Shi¹Jingyi Wang¹Sirui Han⁴Ruoxi Wang¹Linyuan Wang¹Zijing Huan¹Ruiqi Yang¹Desong Hao¹Yanfang Liu¹Xuehua Liu^3*Chao Yuan^1*

• ¹School of Public Health, Tianjin Medical University, Tianjin, China
• ²Nankai University School of Medicine, Tianjin, China
• ³Department of General Practice,Tianjin Union Medical Center, The First Affiliated Hospital of Nankai University, Tianjin, 300121, China., Tianjin, China
• ⁴School of Clinical Medical, Tianjin Medical University, Tianjin, China

Human microbiota is a major factor contributing to the immune system, which offers an opportunity to develop non-invasive methods for disease diagnosis. In some research about Autoimmune Diseases (AIDs), gut microbiota variation has been observed. However, there remains a paucity of research that explores the potential of gut microbiota as a microbial signature for the classification and diagnosis of multi-AIDs. In this research, we analyzed 1,954 gut microbiota sequencing data from public databases, collected from 1,043 patients with 10 AIDs, to identify common or unique microbial signatures for AIDs through differential abundance testing and machine learning techniques. We evaluated five popular algorithms, including RF, SVM, KNN, MLP, and XGBoost models. Five-fold cross-validation and grid search were used to select the model parameters. After comparing the performance of these models, XGBoost model showed superior performance and achieved an area under the receiver operating characteristic curve (AUROC) ranging from 0.75 to 0.99 when predicting different diseases in the test set. At a specificity of 0.7 to 0.96, the sensitivity ranged from 0.66 to 1. Correlating the top 77 microbiota genus with disease phenotypes, 126 significant associations were identified. (false discovery rate [FDR] < 0.05). We improved detection accuracy and disease specificity for AIDs and revealed microbiota features that are specific to 10 different AIDs. Moreover, we found that the changing trends of the shared microbiota features across some AIDs phenotypes are consistent, such as CD and UC. At same time, opposite changing trends of the shared microbial signatures were found, such as Psoriasis and MG. These results suggest that specific gut microbiota genus may affect the host's immunity and induced different AIDs phenotypes. This research holds potential for clinical applications in auxiliary diagnostic evaluation and the monitoring of treatment responses. At the same time, it provides important clues for the research on the characteristics of the intestinal immune microenvironment for different AIDs.