Stacked ensembles on basis of parentage information can predict hybrid performance with an accuracy comparable to marker-based GBLUP

Testcross factorials in newly established hybrid breeding programs are often highly unbalanced, incomplete, and characterized by predominance of special combining ability (SCA) over general combining ability (GCA). This results in a low efficiency of GCA-based selection. Machine learning algorithms might improve prediction of hybrid performance in such testcross factorials, as they have been successfully applied to find complex underlying patterns in sparse data. Our objective was to compare the prediction accuracy of machine learning algorithms to that of GCA-based prediction and genomic best linear unbiased prediction (GBLUP) in six unbalanced incomplete factorials from hybrid breeding programs of rapeseed, wheat, and corn. We investigated a range of machine learning algorithms with three different types of predictor variables: (a) information on parentage of hybrids, (b) in addition hybrid performance of crosses of the parental lines with other crossing partners, and (c) genotypic marker data. In two highly incomplete and unbalanced factorials from rapeseed, in which the SCA variance contributed considerably to the genetic variance, stacked ensembles of gradient boosting machines based on parentage information outperformed GCA prediction. The stacked ensembles increased prediction accuracy from 0.39 to 0.45, and from 0.48 to 0.54 compared to GCA prediction. The prediction accuracy reached by stacked ensembles without marker data reached values comparable to those of GBLUP that requires marker data. We conclude that hybrid prediction with stacked ensembles of gradient boosting machines based on parentage information is a promising approach that is worth further investigations with other data sets in which SCA variance is high.

Efficient Genomic Prediction of Yield and Dry Matter in Hybrid Potato

There is an ongoing endeavor within the potato breeding sector to rapidly adapt potato from a clonal polyploid crop to a diploid hybrid potato crop. While hybrid breeding allows for the efficient generation and selection of parental lines, it also increases breeding program complexity and results in longer breeding cycles. Over the past two decades, genomic prediction has revolutionized hybrid crop breeding through shorter breeding cycles, lower phenotyping costs, and better population improvement, resulting in increased genetic gains for genetically complex traits. In order to accelerate the genetic gains in hybrid potato, the proper implementation of genomic prediction is a crucial milestone in the rapid improvement of this crop. The authors of this paper set out to test genomic prediction in hybrid potato using current genotyped material with two alternative models: one model that predicts the general combining ability effects (GCA) and another which predicts both the general and specific combining ability effects (GCA+SCA). Using a training set comprising 769 hybrids and 456 genotyped parental lines, we found that reasonable a prediction accuracy could be achieved for most phenotypes with both zero common parents (ρ=0.36-0.61) and one (ρ=0.50-0.68) common parent between the training and test sets. There was no benefit with the inclusion of non-additive genetic effects in the GCA+SCA model despite SCA variance contributing between 9% and 19% of the total genetic variance. Genotype-by-environment interactions, while present, did not appear to affect the prediction accuracy, though prediction errors did vary across the trial's targets. These results suggest that genomically estimated breeding values on parental lines are sufficient for hybrid yield prediction.

Modeling the influence of phenotypic plasticity on maize hybrid performance

Phenotypic plasticity, the ability of an individual to alter its phenotype in response to changes in the environment, has been proposed as a target for breeding crop varieties with high environmental fitness. Here, we used phenotypic and genotypic data from multiple maize (Zea mays L.) populations to mathematically model phenotypic plasticity in response to the environment (PPRE) in inbred and hybrid lines. PPRE can be simply described by a linear model in which the two main parameters, intercept a and slope b, reflect two classes of genes responsive to endogenous (class A) and exogenous (class B) signals that coordinate plant development. Together, class A and class B genes contribute to the phenotypic plasticity of an individual in response to the environment. We also made connections between phenotypic plasticity and hybrid performance or general combining ability (GCA) of yield using 30 F₁ hybrid populations generated by crossing the same maternal line with 30 paternal lines from different maize heterotic groups. We show that the parameters a and b from two given parental lines must be concordant to reach an ideal GCA of F₁ yield. We hypothesize that coordinated regulation of the two classes of genes in the F₁ hybrid genome is the basis for high GCA. Based on this theory, we built a series of predictive models to evaluate GCA in silico between parental lines of different heterotic groups.

Development of male sterile lines of CMS chilies (Capsicum annuum L.) from F1 hybrids

Selfing and crossing methods were used to develop the cytoplasmic male sterility (CMS) lines from 2 elite F₁ hybrids of CMS hot chilies. The pungency of the CMS lines was improved by backcrossing with the B cultivar. The first and second backcrossed progenies of the CMS lines showed significantly higher capsaicin contents than the F₁ hybrids. One good female line K16 × BBC2 (K16), was selected and backcrossed with 3 good maintainer cultivars, C5, C9 and C0. Some incomplete male sterility of pollens was demonstrated in the F₁ hybrids and the 1^st backcrossed progenies while the partial sterility disappeared by the stage of the second and third generations of backcrossing. When K16 and P32 were crossed with restorers, fruit yields and yield components of certain F₁ hybrids, parental lines and commercial varieties were significantly different. Heterosis of yield and yield components of the F₁ hybrid chilies was significant. When K16 was used as a female parent, positive and significant heterosis of the F₁ hybrids was the same as P32. Moreover, significant GCA of the restorer lines, C7, C8 and C9, was observed in some horticultural characteristics. Furthermore, significant differences of the specific combining ability of some characteristics were observed in a few F₁ hybrids.

Development of the maize 5.5K loci panel for genomic prediction through genotyping by target sequencing

Genotyping platforms are important for genetic research and molecular breeding. In this study, a low-density genotyping platform containing 5.5K SNP markers was successfully developed in maize using genotyping by target sequencing (GBTS) technology with capture-in-solution. Two maize populations (Pop1 and Pop2) were used to validate the GBTS panel for genetic and molecular breeding studies. Pop1 comprised 942 hybrids derived from 250 inbred lines and four testers, and Pop2 contained 540 hybrids which were generated from 123 new-developed inbred lines and eight testers. The genetic analyses showed that the average polymorphic information content and genetic diversity values ranged from 0.27 to 0.38 in both populations using all filtered genotyping data. The mean missing rate was 1.23% across populations. The Structure and UPGMA tree analyses revealed similar genetic divergences (76-89%) in both populations. Genomic prediction analyses showed that the prediction accuracy of reproducing kernel Hilbert space (RKHS) was slightly lower than that of genomic best linear unbiased prediction (GBLUP) and three Bayesian methods for general combining ability of grain yield per plant and three yield-related traits in both populations, whereas RKHS with additive effects showed superior advantages over the other four methods in Pop1. In Pop1, the GBLUP and three Bayesian methods with additive-dominance model improved the prediction accuracies by 4.89-134.52% for the four traits in comparison to the additive model. In Pop2, the inclusion of dominance did not improve the accuracy in most cases. In general, low accuracies (0.33-0.43) were achieved for general combing ability of the four traits in Pop1, whereas moderate-to-high accuracies (0.52-0.65) were observed in Pop2. For hybrid performance prediction, the accuracies were moderate to high (0.51-0.75) for the four traits in both populations using the additive-dominance model. This study suggests a reliable genotyping platform that can be implemented in genomic selection-assisted breeding to accelerate maize new cultivar development and improvement.

Dissection of Genetic Effects, Heterosis, and Inbreeding Depression for Phytochemical Traits in Coriander

Increasing seed yield, fatty acids, and essential oil content are the main objectives in breeding coriander. However, in order to achieve this, there is a need to understand the nature of gene action and quantify the heterosis and inbreeding depression. Towards this, six genetically diverse parents, their 15 F₁ one-way hybrids, and 15 F₂ populations were evaluated under different water treatments. The genetic effects of general (GCA) and specific combining ability (SCA) and their interactions with water treatment were significant for five traits. Water deficit stress decreased all traits in both F₁ and F₂ generations except for the essential oil content, which was significantly increased due to water deficit stress. Under water deficit stress, a non-additive gene action was predominant in the F₁ generation, while an additive gene action was predominant in the F₂ generation for all the traits except seed yield under severe water deficit stress. There was a positive high heterosis for the traits examined in some hybrids. Furthermore, in the F₂ generation, even after inbreeding depression, some promising populations displayed appropriate mean performance. The results show that the parents used for crossing had a rich, diverse gene pool for the traits studied. Therefore, selection between the individuals of relevant F₂ populations could be used to develop high yielding hybrids or superior lines.

Optimizing predictions in IRRI's rice drought breeding program by leveraging 17 years of historical data and pedigree information

Prediction models based on pedigree and/or molecular marker information are now an inextricable part of the crop breeding programs and have led to increased genetic gains in many crops. Optimization of IRRI's rice drought breeding program is crucial for better implementation of selections based on predictions. Historical datasets with precise and robust pedigree information have been a great resource to help optimize the prediction models in the breeding programs. Here, we leveraged 17 years of historical drought data along with the pedigree information to predict the new lines or environments and dissect the G × E interactions. Seven models ranging from basic to proposed higher advanced models incorporating interactions, and genotypic specific effects were used. These models were tested with three cross-validation schemes (CV1, CV2, and CV0) to assess the predictive ability of tested and untested lines in already observed environments and tested lines in novel or new environments. In general, the highest prediction abilities were obtained when the model accounting interactions between pedigrees (additive) and environment were included. The CV0 scheme (predicting unobserved or novel environments) reveals very low predictive abilities among the three schemes. CV1 and CV2 schemes that borrow information from the target and correlated environments have much higher predictive abilities. Further, predictive ability was lower when predicting lines in non-stress conditions using drought data as training set and/or vice-versa. When predicting the lines using the data sets under the same conditions (stress or non-stress data sets), much better prediction accuracy was obtained. These results provide conclusive evidence that modeling G × E interactions are important in predictions. Thus, considering G × E interactions would help to build enhanced genomic or pedigree-based prediction models in the rice breeding program. Further, it is crucial to borrow the correlated information from other environments to improve prediction accuracy.

Genomic Prediction of Yield Traits in Single-Cross Hybrid Rice (Oryza sativa L.)

Hybrid rice varieties can outyield the best inbred varieties by 15 – 30% with appropriate management. However, hybrid rice requires more inputs and management than inbred rice to realize a yield advantage in high-yielding environments. The development of stress-tolerant hybrid rice with lowered input requirements could increase hybrid rice yield relative to production costs. We used genomic prediction to evaluate the combining abilities of 564 stress-tolerant lines used to develop Green Super Rice with 13 male sterile lines of the International Rice Research Institute for yield-related traits. We also evaluated the performance of their F₁ hybrids. We identified male sterile lines with good combining ability as well as F₁ hybrids with potential further use in product development. For yield per plant, accuracies of genomic predictions of hybrid genetic values ranged from 0.490 to 0.822 in cross-validation if neither parent or up to both parents were included in the training set, and both general and specific combining abilities were modeled. The accuracy of phenotypic selection for hybrid yield per plant was 0.682. The accuracy of genomic predictions of male GCA for yield per plant was 0.241, while the accuracy of phenotypic selection was 0.562. At the observed accuracies, genomic prediction of hybrid genetic value could allow improved identification of high-performing single crosses. In a reciprocal recurrent genomic selection program with an accelerated breeding cycle, observed male GCA genomic prediction accuracies would lead to similar rates of genetic gain as phenotypic selection. It is likely that prediction accuracies of male GCA could be improved further by targeted expansion of the training set. Additionally, we tested the correlation of parental genetic distance with mid-parent heterosis in the phenotyped hybrids. We found the average mid-parent heterosis for yield per plant to be consistent with existing literature values at 32.0%. In the overall population of study, parental genetic distance was significantly negatively correlated with mid-parent heterosis for yield per plant (r = −0.131) and potential yield (r = −0.092), but within female families the correlations were non-significant and near zero. As such, positive parental genetic distance was not reliably associated with positive mid-parent heterosis.

Identifying quantitative trait loci for the general combining ability of yield-relevant traits in maize

Maize is the most important staple crop worldwide. Many of its agronomic traits present with a high level of heterosis. Combining ability was proposed to exploit the rule of heterosis, and general combining ability (GCA) is a crucial measure of parental performance. In this study, a recombinant inbred line population was used to construct testcross populations by crossing with four testers based on North Carolina design II. Six yield-relevant traits were investigated as phenotypic data. GCA effects were estimated for three scenarios based on the heterotic group and the number of tester lines. These estimates were then used to identify quantitative trait loci (QTL) and dissect genetic basis of GCA. A higher heritability of GCA was obtained for each trait. Thus, testing in early generation of breeding may effectively select candidate lines with relatively superior GCA performance. The GCA QTL detected in each scenario was slightly different according to the linkage mapping. Most of the GCA-relevant loci were simultaneously detected in all three datasets. Therefore, the genetic basis of GCA was nearly constant although discrepant inbred lines were appointed as testers. In addition, favorable alleles corresponding to GCA could be pyramided via marker-assisted selection and made available for maize hybrid breeding.

Heterosis Breeding in Eggplant (Solanum melongena L.): Gains and Provocations

Heterosis (or hybrid vigor) results in a hybrid’s phenotypic superiority over its founder parents for quantitative and qualitative traits. Hybrid vigor is defined by mechanisms such as dominant complementation, over-dominance, and epistasis. Eggplant (Solanum melongena L.) is an essential vegetable crop and a good source of dietary minerals, vitamins, and anthocyanins, with a high oxygen radical absorbance capacity and low caloric value. Given the economic and nutritional significance of eggplants, breeding efforts focus on developing high-yielding varieties—mostly F₁ hybrids—with important traits. Studies indicate the successful exploitation of heterosis in the eggplant for a considerable improvement with respect to quantitative traits. In this direction, estimating heterosis for yield-related traits could well be useful for examining the most beneficial hybrid mix with the exploitation of top-quality hybrid. This review examines the current perception of the breeding and molecular aspects of heterosis in eggplants and cites several studies describing the mechanisms. Rendering and combining recent genomics, epigenetic, proteomic, and metabolomics studies present new prospects towards the understanding of the regulatory events of heterosis involved in the evolution and the domestication of the eggplant ideotype.

Enhancing Hybrid Prediction in Pearl Millet Using Genomic and/or Multi-Environment Phenotypic Information of Inbreds

Genomic selection (GS) is an emerging methodology that helps select superior lines among experimental cultivars in plant breeding programs. It offers the opportunity to increase the productivity of cultivars by delivering increased genetic gains and reducing the breeding cycles. This methodology requires inexpensive and sufficiently dense marker information to be successful, and with whole genome sequencing, it has become an important tool in many crops. The recent assembly of the pearl millet genome has made it possible to employ GS models to improve the selection procedure in pearl millet breeding programs. Here, three GS models were implemented and compared using grain yield and dense molecular marker information of pearl millet obtained from two different genotyping platforms (C [conventional GBS RAD-seq] and T [tunable GBS tGBS]). The models were evaluated using three different cross-validation (CV) schemes mimicking real situations that breeders face in breeding programs: CV2 resembles an incomplete field trial, CV1 predicts the performance of untested hybrids, and CV0 predicts the performance of hybrids in unobserved environments. We found that (i) adding phenotypic information of parental inbreds to the calibration sets improved predictive ability, (ii) accounting for genotype-by-environment interaction also increased the performance of the models, and (iii) superior strategies should consider the use of the molecular markers derived from the T platform (tGBS).

Diallel Analysis of Early Leaf Spot (Cercospora arachidicola Hori) Disease Resistance in Groundnut

Early leaf spot (ELS) is one of the major biotic constraints of groundnut production in West and Central Africa. A study using 6 × 6 F2 full diallel populations from six parents (NAMA, B188, PC79-79, QH243C, TS32-1, and CN94C) was conducted to assess the mode of inheritance of ELS resistance traits. The F2 and parents were grown in a randomized complete block design with three replications. Data was collected on ELS disease severity, and an area under disease progress curve (AUDPC) was estimated. The results revealed that additive and non-additive gene actions were involved in the inheritance of the ELS resistance traits, but additive gene action was predominant. Significant reciprocal cross effect was observed, suggesting cytoplasmic effect on ELS resistance. Graphical analysis also revealed the predominance of additive gene action for ELS resistance. The results suggest that early generation selection should be effective for ELS resistance. Looking at the distribution of array points along with the regression line, parental lines NAMA, PC79-79, and B188 would be suitable as good donors in an ELS disease resistance breeding program.

Genome-Wide Single-Nucleotide Polymorphisms in CMS and Restorer Lines Discovered by Genotyping Using Sequencing and Association with Marker-Combining Ability for 12 Yield-Related Traits in Oryza sativa L. subsp. Japonica

Heterosis or hybrid vigor is closely related with general combing ability (GCA) of parents and special combining ability (SCA) of combinations. The evaluation of GCA and SCA facilitate selection of parents and combinations in heterosis breeding. In order to improve combining ability (CA) by molecular marker assist selection, it is necessary to identify marker loci associated with the CA. To identify the single nucleotide polymorphisms (SNP) loci associated with CA in the parental genomes of japonica rice, genome-wide discovered SNP loci were tested for association with the CA of 18 parents for 12 yield-related traits. In this study, 81 hybrids were created and evaluated to calculate the CA of 18 parents. The parents were sequenced by genotyping by sequencing (GBS) method for identification of genome-wide SNPs. The analysis of GBS indicated that the successful mapping of 9.86 × 10⁶ short reads in the Nipponbare reference genome consists of 39,001 SNPs in parental genomes at 11,085 chromosomal positions. The discovered SNPs were non-randomly distributed within and among the 12 chromosomes of rice. Overall, 20.4% (8026) of the discovered SNPs were coding types, and 8.6% (3344) and 9.9% (3951) of the SNPs revealed synonymous and non-synonymous changes, which provide valuable knowledge about the underlying performance of the parents. Furthermore, the associations between SNPs and CA indicated that 362 SNP loci were significantly related to the CA of 12 parental traits. The identified SNP loci of CA in our study were distributed genome wide and caused a positive or negative effect on the CA of traits. For the yield-related traits, such as grain thickness, days to heading, panicle length, grain length and 1000-grain weight, a maximum number of positive SNP loci of CA were found in CMS A171 and in the restorers LC64 and LR27. On an individual basis, some of associated loci that resided on chromosomes 2, 5, 7, 9, and 11 recorded maximum positive values for the CA of traits. From our results, we suggest that heterosis in japonica rice would be improved by pyramiding the favorable SNP loci of CA and eliminating the unfavorable loci from parental genomes.

Genomic Prediction of Single Crosses in the Early Stages of a Maize Hybrid Breeding Pipeline

Prediction of single-cross performance has been a major goal of plant breeders since the beginning of hybrid breeding. Recently, genomic prediction has shown to be a promising approach, but only limited studies have examined the accuracy of predicting single-cross performance. Moreover, no studies have examined the potential of predicting single crosses among random inbreds derived from a series of biparental families, which resembles the structure of germplasm comprising the initial stages of a hybrid maize breeding pipeline. The main objectives of this study were to evaluate the potential of genomic prediction for identifying superior single crosses early in the hybrid breeding pipeline and optimize its application. To accomplish these objectives, we designed and analyzed a novel population of single crosses representing the Iowa Stiff Stalk synthetic/non-Stiff Stalk heterotic pattern commonly used in the development of North American commercial maize hybrids. The performance of single crosses was predicted using parental combining ability and covariance among single crosses. Prediction accuracies were estimated using cross-validation and ranged from 0.28 to 0.77 for grain yield, 0.53 to 0.91 for plant height, and 0.49 to 0.94 for staygreen, depending on the number of tested parents of the single cross and genomic prediction method used. The genomic estimated general and specific combining abilities showed an advantage over genomic covariances among single crosses when one or both parents of the single cross were untested. Overall, our results suggest that genomic prediction of single crosses in the early stages of a hybrid breeding pipeline holds great potential to redesign hybrid breeding and increase its efficiency.

Inheritance of Resistance to Sorghum Shoot Fly, Atherigona soccata in Sorghum, Sorghum bicolor (L.) Moench

Sorghum production is affected by a wide array of biotic constraints, of which sorghum shoot fly, Atherigona soccata is the most important pest, which severely damages the sorghum crop during the seedling stage. Host plant resistance is one of the major components to control sorghum shoot fly, A. soccata. To understand the nature of gene action for inheritance of shoot fly resistance, we evaluated 10 parents, 45 F1's and their reciprocals in replicated trials during the rainy and postrainy seasons. The genotypes ICSV 700, Phule Anuradha, ICSV 25019, PS 35805, IS 2123, IS 2146, and IS 18551 exhibited resistance to shoot fly damage across seasons. Crosses between susceptible parents were preferred for egg laying by the shoot fly females, resulting in a susceptible reaction. ICSV 700, ICSV 25019, PS 35805, IS 2123, IS 2146, and IS 18551 exhibited significant and negative general combining ability (gca) effects for oviposition, deadheart incidence, and overall resistance score. The plant morphological traits associated with expression of resistance/susceptibility to shoot fly damage such as leaf glossiness, plant vigor, and leafsheath pigmentation also showed significant gca effects by these genotypes, suggesting the potential for use as a selection criterion to breed for resistance to shoot fly, A. soccata. ICSV 700, Phule Anuradha, IS 2146 and IS 18551 with significant positive gca effects for trichome density can also be utilized in improving sorghums for shoot fly resistance. The parents involved in hybrids with negative specific combining ability (sca) effects for shoot fly resistance traits can be used in developing sorghum hybrids with adaptation to postrainy season. The significant reciprocal effects of combining abilities for oviposition, leaf glossy score and trichome density suggested the influence of cytoplasmic factors in inheritance of shoot fly resistance. Higher values of variance due to specific combining ability (σ(2)s), dominance variance (σ(2)d), and lower predictability ratios than the variance due to general combining ability (σ(2)g) and additive variance (σ(2)a) for shoot fly resistance traits indicated the predominance of dominance type of gene action, whereas trichome density, leaf glossy score, and plant vigor score with high σ(2)g, additive variance, predictability ratio, and the ratio of general combining ability to the specific combining ability showed predominance of additive type of gene action indicating importance of heterosis breeding followed by simple selection in breeding shoot fly-resistant sorghums. Most of the traits exhibited high broadsense heritability, indicating high inheritance of shoot fly resistance traits.

Genetic Analysis in Maize Foundation Parents with Mapping Population and Testcross Population: Ye478 Carried More Favorable Alleles and Using QTL Information Could Improve Foundation Parents

The development of maize foundation parents is an important part of genetics and breeding research, and applying new genetic information to produce foundation parents has been challenging. In this study, we focused on quantitative trait loci (QTLs) and general combining ability (GCA) of Ye478, a widely used foundation parent in China. We developed three sets of populations for QTL mapping and to analyze the GCA for some agronomic traits. The assessment of 15 traits resulted in the detection of 251 QTLs in six tested environments, with 119 QTLs identified through a joint analysis across all environments. Further, analyses revealed that most favorable alleles for plant type-related traits were from Ye478, and more than half of the favorable alleles for yield-related traits were from R08, another foundation parent used in southwestern China, suggesting that different types of foundation parents carried different favorable alleles. We observed that the GCA for most traits (e.g., plant height and 100-kernel weight) was maintained in the inbred lines descended from the foundation parents. Additionally, the continuous improvement in the GCA of the descendants of the foundation parents was consistent with the main trend in maize breeding programs. We identified three significant genomic regions that were highly conserved in three Ye478 descendants, including the stable QTL for plant height. The GCA for the traits in the F₇ generation revealed that the QTLs for the given traits per se were affected by additive effects in the same way in different populations.