Figure 5.

Testing design for the independent prediction of agronomic traits using the MTME model. Each year, a set of elite and advanced lines is evaluated over multiple locations. The sparse testing design proposes phenotyping of elite lines in all the environments (five in this scenario) and advanced lines in fewer environments (three in this scenario). For independent prediction, the dataset from 2018–19 comprised 55 elite lines with checks and 96 advanced lines. The 2019–20 dataset comprised 42 elite lines with checks and 114 advanced lines. Five environments: BRK, Brookings; DL, Dakota Lakes; HYS, Hayes; OND, Onida; and WIN, Winner.