[Shared mechanisms of obesity and male infertility and prediction of natural therapeutics based on machine learning and single-cell sequencing].

The incidence of obesity and male infertility continues to rise, and a complex pathophysiological association exists between the two. However, the comorbid molecular mechanisms remain incompletely elucidated. In this study, bioinformatics and machine learning methods were integrated to systematically explore the shared mechanisms of obesity and male infertility, and to predict natural compounds and TCMs with therapeutic potential based on multi-omics data. Datasets of male infertility and obesity were obtained from the GEO database, and 43 intersecting genes were identified through differential expression analysis. Using three machine learning algorithms, including random forest, least absolute shrinkage and selection operator(LASSO) regression, and support vector machine(SVM), five Hub genes(MAP4K4, GPT2, ADSSL1, PAIP2, and GINS3) were screened. The combined diagnostic model achieved an area under the curve(AUC) greater than 0.8, indicating good diagnostic performance. Functional enrichment analysis revealed that these genes are mainly involved in key biological processes such as amino acid metabolism, cell migration, and DNA replication. Immune infiltration analysis showed a significant upregulation of central memory CD4~+ T cells in both diseases, suggesting that chronic immune activation is an important basis for their comorbidity. Single-cell sequencing and pseudotime analyses demonstrated disordered cell differentiation trajectories in the adipose tissue of obese patients and the testicular tissue of infertile patients. Based on the connectivity map(CMap) database, ten natural compounds, including berberine, curcumin, and ginsenosides, were predicted and further validated by molecular docking to have strong binding affinities with the Hub genes. Integration with the traditional Chinese medicine systems pharmacology(TCMSP) platform enabled the construction of a "target-natural compound-herbal medicine" interaction network, identifying 38 corresponding herbal medicines. Property and meridian analysis indicated a predominance of warm nature, sweet flavor, and liver meridian tropism, consistent with the TCM therapeutic principles of "resolving phlegm and dampness, strengthening the spleen, and tonifying the kidney". This study reveals, at the molecular level, the comorbid mechanisms of obesity and male infertility in immune-inflammatory regulation and cell differentiation dysfunction, providing a theoretical basis and potential drug targets for precise TCM intervention in obesity-related male infertility.