Freeze-Thaw Imaging for Microorganism Classification Assisted with Artificial Intelligence

Area of Science:

• Nanotechnology
• Microbiology
• Artificial Intelligence

Background:

• Traditional microbial classification methods are often complex, require skilled personnel, and necessitate sophisticated equipment.
• There is a critical need for fast, cost-effective, and accessible microbial identification techniques.

Purpose of the Study:

• To develop and validate a low-cost, user-friendly microbial classification system using gold nanoparticles (AuNPs) and artificial intelligence.
• To demonstrate the efficacy of the freeze-thaw-induced floating pattern of AuNPs (FTFPA) method for identifying various microbes.

Main Methods:

• Coincubation of microbes with AuNPs, followed by freeze-thawing to induce distinct floating patterns.
• Digitization of these patterns to train artificial intelligence models for microbial classification.
• Development of hierarchical classification models for scalability across different taxonomic levels.

Main Results:

• The system achieved a positive sample detection F1 score of 0.976 and a multispecies classification macro F1 score of 0.859.
• Hierarchical classification models demonstrated high performance, with the Enterobacteriales level model reaching a macro F1 score of 0.958.
• The cost per sample for microbial identification was remarkably low at $0.0023.

Conclusions:

• The FTFPA method offers a user-friendly, cost-effective, and convenient platform for microbial identification.
• This AI-coupled nanotechnology approach provides a scalable and accessible solution for clinical diagnosis, environmental monitoring, and food safety applications.