
Journal of Integrative Plant Biology, Journal Year: 2025, Volume and Issue: unknown
Published: April 14, 2025
ABSTRACT Deep learning‐based genomic prediction (DL‐based GP) has shown promising performance compared to traditional GP methods in plant breeding, particularly handling large, complex multi‐omics data sets. However, the effective development and widespread adoption of DL‐based still face substantial challenges, including need for high‐quality sets, inconsistencies benchmarking, integration environmental factors. Here, we summarize key obstacles impeding models propose future developing directions, such as modular approaches, augmentation, advanced attention mechanisms.
Language: Английский