Harnessing high-throughput phenotyping and artificial intelligence for soybean breeding: From trait assessment to...

Area of Science:

• Agricultural Science
• Plant Breeding
• Computational Biology

Background:

• Soybean is a crucial global crop for food, feed, and industrial uses.
• Traditional phenotyping is slow, labor-intensive, and destructive.
• High-Throughput Phenotyping (HTP) offers precise, rapid, and non-destructive trait assessment.

Purpose of the Study:

• To review the application of HTP and AI in soybean breeding.
• To discuss challenges and potential of Deep Learning (DL) in soybean improvement.
• To explore strategies for developing large foundation models for soybean.

Main Methods:

• Review of current literature on HTP platforms and AI applications in soybean.
• Analysis of DL techniques applied to diverse soybean datasets.
• Examination of HTP data integration for breeding.

Main Results:

• HTP enables precise measurement of diverse soybean traits.
• DL shows significant potential for accelerating soybean breeding goals.
• Challenges exist in integrating HTP and AI into existing breeding programs.

Conclusions:

• HTP and AI, particularly DL and Large Language Models (LLMs), can revolutionize soybean breeding.
• Overcoming implementation challenges is key to leveraging these advanced technologies.
• Future research should focus on integrating HTP data and LLMs for enhanced soybean crop development.