Candidate biomarkers of monocyte immunity and metabolism for tuberculosis diagnosis identified using single-cell analysis and explainable AI
Background: Tuberculosis diagnostic clues in host immune and metabolic changes
A six-gene combination reflecting immune and metabolic changes in monocytes has been proposed as a candidate diagnostic biomarker for tuberculosis. The researchers combined single-cell analysis and machine learning to identify candidates, then conducted preliminary validation in independent transcriptomic cohorts and a pilot clinical cohort. The study, published in Archives of Microbiology, focuses on diagnostic candidate discovery and early validation.
The study set out to find markers that reflect abnormalities in host immunity and metabolism while distinguishing tuberculosis rapidly and accurately. The researchers focused on macrophage-related immune and metabolic pathways, using single-cell RNA sequencing (scRNA-seq) as an exploratory tool to examine intercellular signaling in tuberculosis. They then analyzed data from several transcriptomic cohorts to narrow down gene combinations for potential diagnostic use.
Key findings: A six-gene model and preliminary protein-level validation
Single-cell analysis revealed changes in immune states centered on myeloid immune cells. The researchers interpret activated monocytes in particular as key cells organizing inflammatory responses through signals related to cell adhesion and migration.
The researchers evaluated 113 algorithm combinations and selected a model combining glmBoost and random forest (RF). It was based on six genes: IL1B, ATG3, CYBB, MX1, RPS27A, and RPS3. In the abstract, the researchers call this a mitochondria- and macrophage-related signature and interpret it as a combination reflecting changes in monocyte immune and metabolic states in tuberculosis.
The model achieved an area under the receiver operating characteristic curve (AUC) above 0.79 in independent cohorts. AUC measures a model’s ability to distinguish groups; it should not be interpreted as equivalent to 79% accuracy. Shapley Additive exPlanations (SHAP) were used to interpret the model. However, the provided abstract gives no cohort-specific performance, confidence intervals, or figures for comparator models, so whether the model outperforms existing tests remains undetermined.
In an independent pilot clinical cohort, the researchers conducted a prospective evaluation using enzyme-linked immunosorbent assays (ELISAs). They reported elevated protein levels of IL1B, ATG3, CYBB, and MX1 in patients with tuberculosis. These changes were observed at the protein level in addition to the transcriptomic analysis, but elevation of these four proteins alone does not establish the clinical validity of the full six-gene panel.
Meaning and outlook: Distinguish diagnostic candidates from treatment-target exploration
The researchers suggest that the signature could be developed as a triage biomarker to identify people who should receive additional diagnostic testing. The provided abstract does not specify sample size, control-group composition, specimen type, sensitivity, specificity, or test thresholds. Determining who could be tested and how in clinical practice therefore remains a task for future validation.
For treatment-target exploration, the researchers used network pharmacology and molecular docking. The computational results suggest that galangin and kaempferol may have high theoretical binding affinity at binding sites on IL1B and ATG3. This is a prediction to help narrow candidates for follow-up experiments; it does not demonstrate actual target modulation or anti-tuberculosis efficacy. The preliminary performance of the diagnostic model and the computational predictions for treatment candidates need to be assessed separately.
Source: Europe PMC bibliographic record and abstract.
Tuberculosis (TB) remains a formidable global health threat, yet rapid and accurate diagnostic biomarkers capturing host immune-metabolic dysregulation remain elusive. Here, we aimed to map the TB immune microenvironment and engineer a reliable, explainable diagnostic signature targeting the macrophage-associated immune-metabolic axis. Utilizing single-cell RNA sequencing (scRNA-seq) as an exploratory discovery tool, we initially dissected the intercellular communication network in TB. We then deployed an exhaustive consensus machine learning framework, comprising 113 algorithm combinations, across multiple transcriptomic cohorts to pinpoint core diagnostic features. The optimal model was interpreted using SHapley Additive exPlanations (SHAP) and prospectively evaluated by enzyme-linked immunosorbent assay (ELISA) in an independent pilot clinical cohort. Furthermore, network pharmacology and molecular docking were leveraged to identify potential small-molecule modulators. scRNA-seq analysis highlighted a myeloid-biased immune reprogramming, wherein activated monocytes act as key inflammatory orchestrators through adhesion and migration signaling. Our large-scale machine learning screening identified an optimal glmBoost + RF ensemble model underpinned by a 6-gene mitochondrial-macrophage signature (IL1B, ATG3, CYBB, MX1, RPS27A, RPS3), achieving consistent diagnostic discrimination (AUC > 0.79) across independent cohorts. Pilot clinical ELISA validation confirmed the systemic elevation of IL1B, ATG3, CYBB, and MX1 proteins in TB patients. Furthermore, computational molecular docking models suggested that the candidate phytochemicals galangin and kaempferol exhibit strong theoretical binding affinities within the catalytic pockets of IL1B and ATG3. We derived and provided preliminary validation for an AI-based, explainable 6-gene signature reflecting monocyte immune-metabolic reprogramming in TB. This signature not only demonstrates translational potential as a triage
The practical value of this study is that it proposes a candidate marker for selecting people with suspected tuberculosis who may need additional confirmatory testing. Follow-up clinical studies can assess how well the six-gene combination distinguishes tuberculosis in real-world patient groups and whether similar signals occur in other diseases. Its potential use can be judged only after test thresholds, reproducibility, sensitivity, and specificity are established. These findings alone do not establish that the marker can replace existing tests or that it has reduced diagnostic delays.
For drug development, the results may provide a starting point for experimentally testing interactions between IL1B and ATG3 and the candidate compounds. Researchers must first determine whether the docking predictions correspond to actual binding and changes in target function; therapeutic efficacy and safety remain separate questions.