W191Mixed Model to Multiple Harvest Location Trial Applied to Genomic Prediction in Coffea canephora
Mixed Model to Multiple Harvest Location Trial Applied to Genomic Prediction in Coffea canephora
Date: Sunday, January 10, 2016
Time: 5:00 PM
Time: 5:00 PM
Room: Pacific Salon 3
Genomic Selection (GS) has been studied in several crops and has shown potential to increase the rate of genetic gain and reduce the length of the breeding cycle. Despite the relevance, there is a modest number of reports applied to the genus Coffea. Nevertheless, the effective implementation depends on the ability to consider genomic models that represent with adequate reliability the breeding scenario in which the species are inserted. Coffee experimentation, in general, is represented for evaluations in multiple locations and harvests (MET), in order to understand the interaction magnitude and predicting the performance of untested genotypes. Therefore, the main objective of this study was to investigate GS models that accommodate MET modeling. For this, an expansion of the traditional GBLUP was proposed in order to accommodate the interactions in the GS model. Different scenarios that mimic coffee breeding were proposed to assess the predictive ability. In terms of goodness of fit, this approach showed the lowest AIC and BIC values and, consequently, the best goodness of fit. The predictive capacity was measured by cross-validation and, in contrast with the GBLUP, the incorporation of MET modeling showed higher predictive accuracy (on average 10-17% higher) and lower prediction errors. All the genomic analyses were performed using the Genotyping by Sequencing (GBS) approach, which showed a good potential to be used in coffee breeding programs. Thus, in conclusion, the results achieved may be used as a basis for additional studies into the Genus Coffea and expanded for other perennial crops that have a similar experimentation design.