Machine learning-assisted fluorescence sensor array for qualitative and quantitative analysis of pyrethroid...

Area of Science:

• Analytical Chemistry
• Materials Science
• Environmental Science

Background:

• Simultaneous detection of structurally similar pyrethroid pesticides (PPs) in food matrices remains a significant analytical challenge.
• Traditional nanosensing methods struggle with discriminating between PPs due to their high structural resemblance.

Purpose of the Study:

• To develop a sensor array capable of discriminating between structurally similar pyrethroid pesticides (PPs).
• To achieve high-throughput identification and accurate concentration prediction of PPs in complex samples.

Main Methods:

• Construction of sensor arrays using three distinct nanocomposite complexes: rhodamine B-CD@Au, rhodamine 6G-CD@Au, and coumarin 6-CD@Au.
• Utilizing differential receptor/analyte and receptor/dye affinities, alongside non-linear fluorescence-analyte relationships.
• Application of multivariate pattern recognition and machine learning algorithms for data analysis and prediction.

Main Results:

• Successful discrimination of four specific PPs: deltamethrin, fenvalerate, cyfluthrin, and fenpropathrin.
• Achieved 100% classification accuracy for high-throughput identification of PPs in unknown samples.
• Demonstrated high accuracy in predicting PP concentrations using a stepwise prediction strategy combined with machine learning.

Conclusions:

• The developed sensor array offers a robust platform for the simultaneous detection and discrimination of structurally similar pyrethroid pesticides.
• This approach significantly advances the analytical capabilities for pesticide residue monitoring in agricultural products.
• The combination of sensor arrays and machine learning provides a powerful tool for food safety analysis.