Predictive biomarkers for latent Mycobacterium tuberculosis infection

Area of Science:

• Biochemistry
• Immunology
• Machine Learning

Background:

• Tuberculosis (TB) remains a major global infectious disease.
• Current diagnostics struggle to differentiate latent TB infection from healthy states.
• Latent TB infection affects a significant portion of the global population.

Purpose of the Study:

• To develop a diagnostic method for latent tuberculosis infection (LTBI).
• To identify protein biomarkers differentiating healthy, LTBI, and active TB (ATBI) individuals.
• To leverage machine learning for proteomic data analysis.

Main Methods:

• Analysis of publicly available proteomic data from saliva and serum samples.
• Utilized a machine learning approach to identify distinguishing protein profiles.
• Compared data from healthy, LTBI, and ATBI individuals in Ethiopia.

Main Results:

• Identified a six-protein signature: Mast Cell Expressed Membrane Protein-1, Hemopexin, Lamin A/C, Small Proline Rich Protein 2F, Immunoglobulin Kappa Variable 4-1, and Voltage Dependent Anion Channel 2.
• This protein profile accurately differentiates between healthy and latently infected individuals.
• The findings suggest these proteins as potential biomarkers for LTBI diagnosis.

Conclusions:

• A combination of six host proteins can serve as accurate biomarkers for diagnosing latent tuberculosis infection.
• This diagnostic approach could aid in surveillance and prevention of severe TB disease, particularly in high-risk areas.
• Machine learning applied to proteomic data shows promise for infectious disease diagnostics.