Background: In recent years, COVID-19 and tuberculosis have emerged as major infectious diseases, significantly contributing to global mortality as respiratory illnesses. There is increasing evidence of a reciprocal influence between these diseases, exacerbating their incidence, severity, and mortality rates.
Methods: This study involved retrieving COVID-19 and tuberculosis data from the GEO database and identifying common differentially expressed genes. Machine learning techniques, specifically random forest analysis, were applied to pinpoint key genes for diagnosing COVID-19. The Cibersort algorithm was employed to estimate immune cell infiltration in individuals with COVID-19. Additionally, single-cell sequencing was used to study the distribution of VNN1 within immune cells, and molecular docking provided insights into potential drugs targeting these critical prognosis genes.
Results: GMNN, SCD, and FUT7 were identified as robust diagnostic markers for COVID-19 across training and validation datasets. Importantly, VNN1 was associated with the progression of severe COVID-19, showing a strong correlation with clinical indicators and immune cell infiltration. Single-cell sequencing demonstrated a predominant distribution of VNN1 in neutrophils, and molecular docking highlighted potential pharmacological targets for VNN1.
Conclusions: This study enhances our understanding of the shared pathogenic mechanisms underlying tuberculosis and COVID-19, providing essential insights that could improve the diagnosis and treatment of severe COVID-19 cases.
Keywords: COVID-19; VNN1; immune infiltration; machine learning; mechanical ventilation; molecular docking; single-cell sequencing; tuberculosis.
Copyright © 2024 Sun, Wang, Zhou, Liang, Zhang, Li, Han, Fei, Cao and Wang.