Using Rapid Chlorophyll Fluorescence Transients to Classify Vitis Genotypes

Area of Science:

• Plant physiology
• Molecular biology
• Computational biology

Background:

• The Kautsky effect, a rapid polyphasic rise in fluorescence, reflects photochemical apparatus organization.
• This organization is influenced by genotype-environment interactions.
• Classifying plant genotypes is crucial for agriculture and research.

Purpose of the Study:

• To evaluate machine learning techniques for classifying plant genotypes using rapid fluorescence transients.
• To compare the performance of different machine learning algorithms in this classification task.
• To assess the feasibility of using fluorescence transients for rapid Vitis genotype classification.

Main Methods:

• Recording rapid fluorescence induction curves (Kautsky effect) in different Vitis species and cultivars.
• Applying machine learning algorithms: k-nearest neighbors, decision trees, artificial neural networks, and genetic programming.
• Establishing phylogenetic relations using molecular markers.

Main Results:

• Genetic programming (75.3%) and neural networks (71.8%) achieved higher classification success rates than k-nearest neighbors (58.5%) or decision trees (51.6%).
• All tested algorithms significantly outperformed random classification (14% success rate).
• Genetic programming demonstrated slightly superior performance compared to neural networks.

Conclusions:

• Rapid fluorescence transients, analyzed with machine learning, are effective for classifying Vitis genotypes.
• Genetic programming shows particular promise for rapid, preliminary classification of Vitis species and cultivars.
• This approach offers a feasible method for high-throughput plant genotype identification.