Identification of monocyte-associated genes MSRB2, CLEC4D, and ASGR2 as potential biomarkers for tuberculosis via machine learning and mendelian randomization.

OBJECTIVE

This study explores the link between immune cells and tuberculosis (TB) pathogenesis and progression, proposing diagnostic strategies based on immune microenvironment changes.

METHODS

The CIBERSORT algorithm assessed immune cell infiltration in TB tissues, validated by routine blood tests. Differential expression analysis and WGCNA identified key genes and modules. GO and KEGG analyses elucidated biological functions. Machine learning pinpointed diagnostic biomarkers and built a predictive model. Further validation included GSVA, single-cell data, Mendelian randomization, and RT-qPCR.

RESULTS

Analysis of the immune microenvironment in TB patients and healthy controls revealed monocytes as the predominant immune cell type. A total of 90 overlapping genes were identified through differential expression analysis and WGCNA. A diagnostic model incorporating MSRB2, CLEC4D, and ASGR2 was constructed using three distinct machine learning algorithms and logistic regression. Single-cell data analysis demonstrated that these three genes were predominantly expressed in mononuclear cells of TB patients. MR analysis further established a causal relationship between CLEC4D and an elevated risk of TB.

CONCLUSION

We established a monocyte-based diagnostic model demonstrating robust predictive accuracy. MSRB2, CLEC4D, and ASGR2 represent promising therapeutic targets for TB immunotherapy, providing potential breakthroughs in diagnostic precision and treatment efficacy.