A new method for detecting mixed bacteria based on multi-wavelength transmission spectroscopy technology

Area of Science:

• Spectroscopy
• Microbiology
• Data Analysis

Background:

• Previous work focused on identifying single bacterial species using multi-wavelength transmission spectra.
• Current research addresses the challenge of spectral analysis for mixed bacterial samples.

Purpose of the Study:

• To develop a method for spectral separation and concentration determination of mixed bacteria.
• To advance the detection and online monitoring of waterborne microbial contamination.

Main Methods:

• Developed a Principal Component Analysis-Monte Carlo (PCA-MC) model for spectral separation of mixed bacteria.
• Utilized a neural network concentration inversion model with separated spectra as input.
• Validated the method with mixtures of S. aureus, K. pneumoniae, and S. typhimurium.

Main Results:

• Achieved low mean relative errors in component analysis (2-6.1%) for various bacterial mixtures.
• Obtained high coefficients of determination (R² > 0.9947) for the concentration inversion model.
• Demonstrated rapid and accurate determination of component ratios and concentrations in mixed bacteria.

Conclusions:

• The proposed PCA-MC and neural network approach effectively separates mixed bacterial spectra and quantifies individual concentrations.
• This method represents a significant advancement for real-time monitoring of waterborne microbial contamination.
• The technique offers a rapid and accurate solution for analyzing complex bacterial mixtures.