Improving Genomic Prediction in Cassava Field Experiments by Accounting for Interplot Competition

Area of Science:

• Agricultural Science
• Plant Breeding
• Genetics

Background:

• Resource competition is inherent in crop cultivation.
• Understanding genotypic competitive ability is vital for plant breeding and commercialization strategies.
• Cassava (Manihot esculenta) competition patterns require detailed investigation.

Purpose of the Study:

• To investigate competition patterns in cassava at both genotypic (interclonal) and plot (interplot competition/error) levels.
• To develop and evaluate genomic selection (GS) models that incorporate competitive abilities.
• To enhance the accuracy of predicting genotypic performance by accounting for competition.

Main Methods:

• Modified incidence matrices to quantify neighbor effects on target genotypes.
• Integrated competitive components into genomic selection models.
• Utilized 10-fold cross-validation repeated five times to assess model predictability.
• Compared model performance using prediction root mean squared error (pRMSE).

Main Results:

• GS models incorporating interclonal competition showed <10% accuracy increase.
• GS models accounting for competition error achieved up to 25% accuracy increase.
• Competitive influence extended beyond immediate neighbors.
• Simulations indicated a 26% accuracy gain in estimating genotypic effects with competitive variance.

Conclusions:

• Accounting for plot-level competition error in GS models substantially improves prediction accuracy in cassava.
• Competitive effects in cassava are non-local, influencing distant genotypes.
• The developed GS models offer a more accurate approach to predicting cassava performance by integrating competitive abilities.