Taxonomic classification of phytoplankton with multivariate optical computing, part III: demonstration

Area of Science:

• Marine biology
• Optical instrumentation
• Biophotonics

Background:

• Phytoplankton identification is crucial for marine ecosystem monitoring.
• Existing methods for phytoplankton analysis can be time-consuming and labor-intensive.
• Fluorescence imaging offers a promising avenue for rapid, high-throughput analysis.

Purpose of the Study:

• To develop and validate an automated algorithm for analyzing fluorescence tracks of phytoplankton.
• To enable precise classification of phytoplankton species using fluorescence intensity ratios.
• To assess the algorithm's performance with known phytoplankton species.

Main Methods:

• Utilized a fluorescence imaging photometer to record phytoplankton fluorescence tracks.
• Developed an algorithm to isolate single phytoplankter transits and identify fluorescence streaks.
• Integrated fluorescence intensity of streaks to calculate ratios for classification.
• Tested the algorithm with 853 fluorescence measurements of Emiliania huxleyi and Thalassiosira pseudonana.

Main Results:

• The algorithm successfully isolated and analyzed fluorescence tracks.
• Calculated fluorescence intensity ratios closely matched theoretical predictions.
• The distribution of ratios for each species was consistent with signal-to-noise ratio calculations.
• Achieved perfect classification with no overlap between the two tested species.

Conclusions:

• The automated analysis of fluorescence tracks provides a robust method for phytoplankton classification.
• This approach offers high accuracy and efficiency for distinguishing between phytoplankton species.
• The findings support the use of fluorescence imaging for advanced phytoplankton monitoring and research.