The use of cross prediction methods in a practical potato breeding programme

Area of Science:

• Plant breeding
• Quantitative genetics
• Horticultural science

Background:

• Traditional cross prediction studies often use limited data, insufficient for large-scale breeding programs.
• Evaluating progeny performance across multiple generations is resource-intensive in potato breeding.

Purpose of the Study:

• To assess the feasibility of cross prediction methods in a practical potato (Solanum tuberosum) breeding scheme.
• To identify reliable early-stage indicators for predicting progeny performance in later generations.

Main Methods:

• Analysis of progeny from 52 potato crosses.
• Utilizing breeder's preference scores (visual assessment of tuber commercial potential) as the primary variate.
• Comparing prediction accuracy based on mean scores, expected proportion of desirable clones, and observed frequency of desirable clones.

Main Results:

• Superior potato crosses could be identified using prediction methods.
• Mean preference scores from seedlings or first-year clonal plants provided the best prediction of third-generation performance.
• Predictions based on the expected proportion of clones exceeding a target value were also effective.
• Observed frequency of desirable clones in early samples yielded the poorest prediction accuracy.

Conclusions:

• Cross prediction methods are viable for potato breeding schemes.
• Early-stage breeder's preference scores offer a reliable and efficient method for identifying high-potential crosses.
• These findings can streamline potato breeding by reducing the need for extensive multi-generational evaluations.