Volumetric Deep Learning-Based Precision Phenotyping of Gene-Edited Tomato for Vertical Farming

Area of Science:

• Agricultural Science
• Genetics and Genomics
• Artificial Intelligence in Agriculture

Background:

• Climate change and urbanization challenge sustainable agriculture and food production.
• Vertical farming offers high-density cultivation but requires precise control of crop height.
• Tomato cultivation needs determinate varieties for efficient vertical farming systems.

Purpose of the Study:

• To develop a new tomato cultivar optimized for vertical farming by editing Gibberellin 20-oxidase (SlGA20ox) genes using CRISPR-Cas9.
• To propose a volumetric model for non-destructive identification of gene-edited mutants via chlorophyll fluorescence analysis.
• To enhance AI-based phenotyping for accelerated crop breeding and genetic resource management.

Main Methods:

• CRISPR-Cas9 gene editing of SlGA20ox genes in tomato to create determinate growth mutants.
• Development of a novel volumetric model for analyzing chlorophyll fluorescence data.
• Application of a deep learning framework for automated 3D feature extraction and temporal analysis of fluorescence imaging data.

Main Results:

• Successfully developed gene-edited tomato plants with suppressed indeterminate growth.
• The proposed volumetric model achieved over 84% classification accuracy in distinguishing gene-edited plants.
• The deep learning framework automated feature extraction and analysis of temporal fluorescence data, outperforming traditional methods.

Conclusions:

• The study demonstrates a successful approach to developing and identifying tomato plants tailored for vertical farming.
• The AI-driven phenotyping method using chlorophyll fluorescence offers a powerful tool for crop breeding.
• This methodology can be extended to other crops and traits, accelerating agricultural innovation.