The digital orchard: advanced data-driven technologies in apple breeding and genetic modification.

Area of Science:

• Horticultural Science
• Plant Breeding
• Agricultural Technology

Background:

• Apple breeding faces challenges due to climate change, pathogens, and slow traditional methods.
• Long juvenile periods and high heterozygosity in apples hinder rapid trait selection.
• Consumer demand for novel traits necessitates faster breeding programs.

Purpose of the Study:

• To systematically review advanced data-driven technologies for accelerating apple breeding and genetic modification.
• To synthesize the current state of digital breeding in Malus × domestica.
• To identify research gaps and future directions in data-driven apple improvement.

Main Methods:

• Systematic literature review (SLR) following the PRISMA-EcoEvo protocol.
• Analysis of 47 selected studies from major scientific databases (Web of Science, Scopus, PubMed).
• Thematic synthesis of findings related to high-throughput phenotyping, machine learning, and genome editing.

Main Results:

• A paradigm shift towards 'digital breeding' integrating high-throughput phenotyping (HTP), machine learning (ML)/deep learning (DL), and genome editing (CRISPR/Cas9).
• HTP automates trait data collection; ML/DL achieves >96% accuracy in cultivar identification and boosts genomic selection predictive ability by 18%.
• CRISPR/Cas9 enables rapid introduction of traits like disease resistance and improved shelf life, with transgene-free methods accelerating commercialization.

Conclusions:

• The convergence of HTP, ML/DL, and genome editing, integrated via agricultural IoT (AIoT), is revolutionizing apple breeding speed and precision.
• Emerging frontiers include federated learning, explainable AI (XAI), and adapting to new regulatory frameworks.
• Standardized datasets and end-to-end system validation are critical research needs for future advancements.