Interactive Hybrid Rice Breeding with Parametric Dual Projection.

Area of Science:

• Agricultural Science
• Genetics
• Bioinformatics

Background:

• Hybrid rice breeding relies on selecting desirable agronomic traits for higher yields.
• Genomic selection enhances hybrid rice breeding by predicting traits from genes, reducing field cultivation workload.
• Current genomic prediction models have limited accuracy, necessitating breeder experience for regulatory gene identification and hybrid selection, a time-consuming process.

Purpose of the Study:

• To develop a visual analysis method for interactive hybrid rice breeding.
• To facilitate the dual-analysis task of regulatory gene identification and hybrid selection.
• To improve the efficiency and accuracy of identifying key genes and selecting superior hybrids.

Main Methods:

• Proposed a visual analysis method for interactive hybrid rice breeding.
• Developed a parametric dual projection method with theoretical guarantees for dual analysis.
• Created gene and hybrid visualization tools based on the dual projection method.

Main Results:

• The parametric dual projection method was quantitatively evaluated.
• Identified key regulatory genes controlling desirable traits in hybrid rice.
• Successfully selected desired hybrids, validated through case studies and breeder feedback.
• Demonstrated the effectiveness of the visual analysis method in a real-world breeding scenario.

Conclusions:

• The proposed visual analysis method significantly facilitates interactive hybrid rice breeding.
• The parametric dual projection and visualization tools enhance the identification of regulatory genes and selection of hybrids.
• This approach offers a more efficient and potentially accurate alternative to traditional methods, reducing breeder workload.