Separation-free bacterial identification in arbitrary media via deep neural network-based SERS analysis

Area of Science:

• Analytical Chemistry
• Biotechnology
• Machine Learning

Background:

• Surface-enhanced Raman spectroscopy (SERS) shows promise for microorganism detection.
• Current SERS methods face challenges with interfering signals and require extensive sample preparation.

Purpose of the Study:

• To develop a faster and simpler method for bacterial detection using SERS.
• To overcome limitations of traditional SERS by eliminating the need for bacterial separation.

Main Methods:

• Utilized SERS for bacterial signal acquisition.
• Developed and applied a novel deep learning model, DualWKNet, for signal classification.
• Tested the system on common bacteria, Escherichia coli (E. coli) and Staphylococcus epidermidis (S. epidermidis).

Main Results:

• Achieved high classification accuracies of up to 98% for bacteria in various media.
• Demonstrated successful

Conclusions:

• The synergistic combination of SERS and DualWKNet offers an effective platform for rapid, separation-free bacterial detection.
• This approach significantly reduces data acquisition time and training data requirements.