Genomic prediction is expected to considerably increase genetic gains by increasing selection intensity and accelerating the breeding cycle. In this study, marker effects estimated in 255 diverse maize (Zea mays L.) hybrids were used to predict grain yield, anthesis date, and anthesis-silking interval within the diversity panel and testcross progenies of 30 F2-derived lines from each of five populations. Although up to 25% of the genetic variance could be explained by cross validation within the diversity panel, the prediction of testcross performance of F2-derived lines using marker effects estimated in the diversity panel was on average zero. Hybrids in the diversity panel could be grouped into eight breeding populations differing in mean performance. When performance was predicted separately for each breeding population on the basis of marker effects estimated in the other populations, predictive ability was low (i.e., 0.12 for grain yield). These results suggest that prediction resulted mostly from differences in mean performance of the breeding populations and less from the relationship between the training and validation sets or linkage disequilibrium with causal variants underlying the predicted traits. Potential uses for genomic prediction in maize hybrid breeding are discussed emphasizing the need of (1) a clear definition of the breeding scenario in which genomic prediction should be applied (i.e., prediction among or within populations), (2) a detailed analysis of the population structure before performing cross validation, and (3) larger training sets with strong genetic relationship to the validation set.

Genotyping-by-sequencing (GBS) technologies have proven capacity for delivering large numbers of marker genotypes with potentially less ascertainment bias than standard single nucleotide polymorphism (SNP) arrays. Therefore, GBS has become an attractive alternative technology for genomic selection. However, the use of GBS data poses important challenges, and the accuracy of genomic prediction using GBS is currently undergoing investigation in several crops, including maize, wheat, and cassava. The main objective of this study was to evaluate various methods for incorporating GBS information and compare them with pedigree models for predicting genetic values of lines from two maize populations evaluated for different traits measured in different environments (experiments 1 and 2). Given that GBS data come with a large percentage of uncalled genotypes, we evaluated methods using nonimputed, imputed, and GBS-inferred haplotypes of different lengths (short or long). GBS and pedigree data were incorporated into statistical models using either the genomic best linear unbiased predictors (GBLUP) or the reproducing kernel Hilbert spaces (RKHS) regressions, and prediction accuracy was quantified using cross-validation methods. The following results were found: relative to pedigree or marker-only models, there were consistent gains in prediction accuracy by combining pedigree and GBS data; there was increased predictive ability when using imputed or nonimputed GBS data over inferred haplotype in experiment 1, or nonimputed GBS and information-based imputed short and long haplotypes, as compared to the other methods in experiment 2; the level of prediction accuracy achieved using GBS data in experiment 2 is comparable to those reported by previous authors who analyzed this data set using SNP arrays; and GBLUP and RKHS models with pedigree with nonimputed and imputed GBS data provided the best prediction correlations for the three traits in experiment 1, whereas for experiment 2 RKHS provided slightly better prediction than GBLUP for drought-stressed environments, and both models provided similar predictions in well-watered environments.

Genomic selection offers great potential for increasing the rate of genetic improvement in plant breeding programs. This research used simulation to evaluate the effectiveness of different strategies for genotyping and phenotyping to enable genomic selection in early generation individuals (e.g., F2) in breeding programs involving biparental or similar (e.g., backcross or top cross) populations. By using phenotypes that were previously collected in other biparental populations, selection decisions could be made without waiting for phenotypes that pertain directly to the selection candidate to be collected, a process that would take at least three growing seasons. If these phenotypes were collected in biparental populations that were closely related to the selection candidates, only a small number of markers (e.g., 200–500) and a small number of phenotypes (e.g., 1000) were needed to achieve effective accuracy of estimated breeding values. If these phenotypes were collected in biparental populations that were not closely related to the selection candidates, as many as 10,000 markers and 5000 to 20,000 phenotypes were needed. Increasing marker density beyond 10,000 markers did not show benefit and in some scenarios reduced the accuracy of prediction. This study provides a guide, enabling resource allocation to be optimized between genotyping and phenotyping investment dependent on the population under development.

To develop stable and high‐yielding maize (Zea mays L.) hybrids for a diverse target population of environments (TPE), breeders have to decide whether greater gains result from selection across the undivided TPE or within more homogeneous subregions. Currently, CIMMYT subdivides the TPE in eastern and southern Africa into climatic and geographic subregions. To study the extent of specific adaptation to these subregions and to determine whether selection within subregions results in greater gains than selection across the undivided TPE, yield data of 448 maize hybrids evaluated in 513 trials across 17 countries from 2001 to 2009 were used. The trials were grouped according to five subdivision systems into climate, altitude, geographic, country, and yield‐level subregions. For the first four subdivision systems, genotype × subregion interaction was low, suggesting broad adaptation of maize hybrids across eastern and southern Africa. In contrast, genotype × yield‐level interactions and moderate genotypic correlations between low‐ and high‐yielding subregions were observed. Therefore, hybrid means should be estimated by stratifying the TPE considering the yield‐level effect as fixed and appropriately weighting information from both subregions. This strategy was at least 10% better in terms of predicted gains than direct selection using only data from the low‐ or high‐yielding subregion and should facilitate the identification of hybrids that perform well in both subregions.

The exchange of elite breeding materials across regions is an important way in which multinational maize breeding programmes access new genetic variation, improve efficiency and reduce costs. Our objectives were to examine whether CIMMYT's breeding programmes for tropical and subtropical environments in Mexico and Eastern and Southern Africa (ESA) can effectively share materials. Sets of selected and unselected lines were evaluated for per se and testcross performance in multiple environments in Mexico and ESA for grain yield, days to anthesis and plant height. Genotypic correlations between performance in Mexico and ESA as testcross and line per se were high (≥ 0.72) for all experiments, and indirect selection efficiency ranged from 67 to over 100% for all traits. Lines selected in ESA or Latin America performed equally well in each region, indicating selection was for broad rather than regional adaptation. Thus, breeding programmes of CIMMYT in both Mexico and ESA can benefit tremendously by exchanging breeding materials and test results, and elite selections from each region should be fast‐tracked for evaluation in the other.