We identified and analyzed two independent publicly available RNA-seq datasets from the Gene Expression Omnibus (GEO) database focusing on peripheral blood samples from NTM patients:

Dataset selection criteria included: (1) peripheral blood samples from NTM patients; (2) appropriate control groups; (3) high-quality RNA-seq or microarray data; and (4) sufficient sample sizes for statistical analysis. Both datasets utilized RNA-seq technology on whole blood samples and provided comprehensive transcriptomic data suitable for comparative analysis. The complete analytical workflow is shown in Figure 1 .

We employed a streamlined bioinformatics pipeline for cross-dataset analysis:

All statistical analyses incorporated appropriate multiple testing corrections (Benjamini-Hochberg FDR < 0.05).

The analyzed datasets comprised a total of 65 whole blood samples across two independent cohorts, providing robust statistical power for cross-dataset analysis. Quality control metrics indicated high-quality data suitable for downstream analysis, with clear separation between NTM and control groups in principal component analysis.

Differential expression analysis revealed distinct transcriptomic signatures between NTM patients and controls ( Figure 2 ). GSE97298 demonstrated broader transcriptomic changes, while GSE290289 showed a more focused response pattern. Functional pathway enrichment analysis revealed predominant dysregulation in immune system processes, cytokine signaling, T-cell activation, and pathogen recognition pathways ( Figure 3 ).

Despite methodological differences between studies, we identified 10 genes commonly dysregulated across both datasets, representing core molecular changes in NTM disease detectable in peripheral blood ( Figure 4 ). The identification of common dysregulated genes across independent datasets provides robust evidence for core pathogenic mechanisms in NTM disease, demonstrating the value of cross-dataset validation approaches for biomarker identification.

The protein-protein interaction network analysis revealed highly coordinated immune dysregulation patterns ( Figure 5 ). The analysis revealed a highly connected network comprising 10 nodes and 12 edges, with a network density of 0.267 indicating efficient information flow. Key topological features included:

The network topology characteristics suggest a highly coordinated immune response dysregulation in NTM disease, with efficient communication pathways between key regulatory nodes ( Figure 6 ).

Three genes emerged as major network hubs based on centrality analysis:

CD36 demonstrated the highest degree centrality (0.556) and betweenness centrality (0.722), positioning it as the most critical node in the network. Interestingly, CD36 showed opposite regulation patterns between the two datasets:

CD36, a scavenger receptor involved in lipid metabolism and pathogen recognition, plays crucial roles in mycobacterial infections through its involvement in fatty acid oxidation and inflammatory responses. The differential regulation may reflect different patient populations, disease stages, or therapeutic interventions between studies.

CD3E exhibited high connectivity (degree centrality: 0.444) and was consistently downregulated across both datasets, indicating compromised T-cell receptor signaling in NTM disease. This finding aligns with previous observations of impaired T-cell responses in NTM patients.

GZMK showed significant connectivity and consistent downregulation across both cohorts, suggesting impaired cytotoxic T-cell and NK cell function. This finding is consistent with the known role of cytotoxic cells in mycobacterial clearance.

The comprehensive analysis reveals several key aspects of CD36’s role in NTM disease ( Figure 7 ).

The opposite regulation of CD36 between datasets suggests molecular heterogeneity in NTM disease, potentially reflecting:

CD36’s central network position and functional diversity make it an attractive candidate for precision medicine applications in NTM disease:

This study provides the first comprehensive cross-dataset transcriptomic analysis of NTM disease using peripheral blood samples, revealing previously unrecognized immune regulatory networks. Our findings demonstrate that despite using different patient cohorts and methodologies, consistent patterns of immune dysfunction can be identified, supporting the existence of core pathogenic mechanisms in NTM disease.

The identification of a 10-gene regulatory network with high connectivity (density = 0.267) suggests coordinated dysregulation of immune responses in NTM disease. This network-based approach provides new insights into the systemic nature of immune dysfunction in NTM infections.

The identification of CD36 as the central network hub represents a significant finding with multiple implications for understanding NTM pathogenesis. CD36’s role as a scavenger receptor involved in lipid metabolism, pathogen recognition, and inflammatory responses positions it at the intersection of multiple pathways critical for host-pathogen interactions.

The differential CD36 regulation between datasets may reflect fundamental aspects of NTM disease heterogeneity:

The consistent downregulation of CD3E and GZMK across both datasets provides strong evidence for adaptive immune dysfunction in NTM disease. This finding aligns with previous observations by Cowman et al., who reported widespread downregulation of T-cell signaling pathways in pulmonary NTM patients [15] .

The implications of adaptive immune suppression extend beyond immediate pathogen clearance difficulties:

The expression patterns identified in this study offer multiple opportunities for biomarker development:

Our findings identify multiple potential therapeutic targets:

Several limitations should be acknowledged:

Future research directions should include:

This study benefits from recent advances in computational biology and systems medicine approaches. The integration of machine learning algorithms [24] and multiple-criteria decision-making frameworks [25] represents a growing trend in precision medicine research. Additionally, emerging therapeutic approaches including bacteriophage therapy [26] and hostdirected therapy targeting ferroptosis pathways [27] may complement the transcriptomic insights identified in this study.