Abstract
Potato is a versatile food crop and major component of human nutrition worldwide. Model calculations and computer simulations can be used to optimize the resource allocation in potato breeding programs but require quantitative genetic parameters. The objectives of our study are to (i) estimate quantitative genetic parameters of the most important phenotypic traits in potato breeding programs, (ii) compare the importance of inter- vs. intra-population variance, (iii) quantify genotypic and phenotypic covariances among phenotypic traits, and (iv) examine the effect of a preselection in the single hills stage on variance and covariance components in later stages of the breeding program. Our study was based on a total of 1066 clones from three breeding programs which were evaluated in a non-orthogonal way in 15 environments for a total of 26 phenotypic traits. The examined traits showed an overall high to medium heritability, and variance analysis revealed trait-specific differences in the influence of the genotypic, environmental, and genotype-environment interaction effect. Accounting for heterogeneity in the residual variances between the 15 environments led to a significant improvement of the variance parameter estimation. The result of our study suggested that the first selection step at the single hills stage did not negatively impact the genetic variability of the target traits implying that the traits assessed in the earlier stages were not correlated with the traits influencing market success. Our results can be used as base for further simulation studies and, thus, help to optimize the resource allocation in breeding programs.