Integrative multi-omics analysis identifies a robust 14-metabolite signature and reveals microbiome-metabolite-host
Leiyang Dai1, Xiao Wang1, Hui Zhang1
1Yunnan Key Laboratory of Laboratory Medicine, Yunnan Clinical Research Center for Laboratory Medicine, Department of Clinical Laboratory, the First Affiliated Hospital of Kunming Medical University, Kunming, Yunnan, China.
Background:
Atherosclerosis (AS) is a complex metabolic and inflammatory disease in which interactions between host metabolism and gut microbiota play critical roles. However, robust metabolic biomarkers and their integration with microbial and host factors remain incompletely understood.
Methods:
We performed untargeted metabolomics to characterize metabolic alterations between AS patients and healthy controls (HC). Differential metabolites were identified and subjected to pathway enrichment analysis. Three machine learning models, including random forest (RF), least absolute shrinkage and selection operator (LASSO), and support vector machine (SVM), were applied to identify key metabolite signatures. Gut microbiota composition was analyzed using 16S rRNA sequencing, and correlation analyses were conducted to explore microbiome-metabolite interactions. In addition, inflammatory and senescence-related markers were assessed to evaluate host responses.
Results:
A total of 122 differential metabolites were identified between AS and HC, primarily enriched in amino acid-related pathways, including tryptophan, phenylalanine, and methionine metabolism. Machine learning integration revealed a robust panel of 14 overlapping metabolites with strong discriminative performance. Among them, Trimethylamine N-oxide, 3-Hydroxyhippuric acid, and Cholesteryl sulfate showed the highest diagnostic potential. Despite limited differences in gut microbial composition, several microbiota-derived metabolites and significant correlations between specific genera and metabolites were observed, suggesting functional alterations in the microbiome. Furthermore, senescence markers P16 and P21 were significantly elevated in AS and were associated with key metabolites and microbial taxa, whereas classical inflammatory markers showed no significant differences.
Conclusion:
This study identifies a robust metabolite signature associated with AS and highlights a coordinated microbiome-metabolite-host interaction network. These findings provide new insights into the metabolic mechanisms underlying AS and suggest potential biomarkers and therapeutic targets for disease diagnosis and intervention.
