Graphene-Based Refractive Index Sensor Using Machine Learning for Detection of Mycobacterium Tuberculosis Bacteria

Area of Science:

• Biosensing
• Nanotechnology
• Machine Learning

Background:

• Rapid detection of Mycobacterium tuberculosis is crucial for reducing tuberculosis (TB) disease burden.
• Existing diagnostic methods can be time-consuming, necessitating faster detection techniques.
• Label-free biosensors offer advantages in simplifying detection processes.

Purpose of the Study:

• To develop a highly sensitive, label-free graphene-based refractive index sensor for detecting Mycobacterium tuberculosis.
• To utilize a machine learning approach for enhanced prediction and analysis of biosensor performance.
• To optimize sensor design parameters for improved sensitivity and diagnostic accuracy.

Main Methods:

• Design of a graphene-SiO2 hybrid layer biosensor with a plus-shaped metasurface for enhanced sensitivity.
• Analysis of key parameters including substrate thickness, resonator thickness, and angle of incidence.
• Application of a K-Nearest Neighbors (KNN)-regressor machine learning model to predict absorption values for wavelength analysis.
• Experimental validation using R2 Score to evaluate the KNN-regressor model's prediction efficiency.

Main Results:

• The proposed graphene-based sensor demonstrates high sensitivity for Mycobacterium tuberculosis detection.
• Machine learning, specifically the KNN-regressor model, effectively predicts absorption values for intermediate wavelengths.
• Lower values of K in the KNN-regressor model yielded high prediction efficiency.
• Comparative analysis indicates the proposed sensor's potential against other published designs.

Conclusions:

• The developed label-free graphene biosensor offers a promising tool for rapid Mycobacterium tuberculosis detection.
• The integration of machine learning significantly enhances the sensor's predictive capabilities for diagnostic applications.
• The sensor's high sensitivity and efficient detection mechanism are suitable for integration into biomedical devices for tuberculosis diagnostics.