Pollen discrimination and classification by Fourier transform infrared (FT-IR) microspectroscopy and machine learning

Area of Science:

• Analytical Chemistry
• Environmental Science
• Immunology

Background:

• Aerobiological monitoring networks typically measure outdoor pollen concentrations.
• Accurate identification of airborne pollen is crucial for allergy management.
• Existing methods for pollen classification can be time-consuming and require expertise.

Purpose of the Study:

• To evaluate mid-infrared Fourier transform infrared (FT-IR) microspectroscopy for pollen discrimination and classification.
• To assess the effectiveness of unsupervised and supervised multivariate statistical methods in analyzing FT-IR pollen spectra.
• To determine the feasibility of applying FT-IR microspectroscopy in aerobiological monitoring.

Main Methods:

• Pollen samples from 11 different taxa were analyzed using FT-IR microspectroscopy.
• Unsupervised hierarchical cluster analysis was employed to assess spectral reproducibility.
• Supervised learning, specifically K-nearest neighbors with leave-one-out cross-validation, was used for classification.
• Analysis was performed on single pollen grain spectra and spectra from groups of grains.

Main Results:

• Hierarchical cluster analysis demonstrated reproducibility of FT-IR spectra for individual pollen grains and groups.
• The K-nearest neighbors classifier achieved an overall accuracy of 84% for pollen classification using single-grain spectra.
• FT-IR microspectroscopy proved to be a reliable method for distinguishing between different pollen types.
• The study discussed the practical limitations of implementing this method in aerobiological stations.

Conclusions:

• FT-IR microspectroscopy, combined with multivariate statistics, offers a robust approach for identifying allergenic pollen.
• The method provides a high degree of accuracy in pollen discrimination and classification.
• This technique has the potential to enhance the capabilities of current aerobiological monitoring systems for airborne allergens.