Acca sellowiana is a subtropical tree in the myrtle family (Myrtaceae) native to southern Brazil and northeastern Uruguay. It is recognized for its value as a fruit-bearing, ornamental, and medicinal species. Based on distinctive characteristics of fruits, seeds, and leaves, as well as its geographical distribution pattern, two variants of the species are distinguished: the “Brazilian type” and the “Uruguayan type”. The objective of this study was to characterize, for the first time, the diversity of 202 individuals from four wild populations in Uruguay, representative of the species’ most southern natural distribution. Twenty-three morphological descriptors (leaf, flower, and fruit) and 204 RAPD (Random Amplified Polymorphic DNA) markers were used. The morphological data collected validated the main criteria that distinguish “Uruguayan type” populations from “Brazilian type” populations, such as lower seed weight and fruit size, thin and slightly rough skin, high pulp percentage, and hairy white abaxial leaf surfaces. Analyses of both morphological and molecular data indicated wide diversity and strong population structuring, which is relevant information for designing conservation plans, sustainable utilization, and genetic improvement of the plant genetic resources of this species.

Background: Single-step genomic BLUP (ssGBLUP) has gained increasing interest from forest tree breeders. ssGBLUP combines phenotypic and pedigree data with marker data to enhance the prediction accuracy of estimated breeding values. However, potential errors in determining progeny relationships among open-pollinated species may result in lower accuracy of estimated breeding values. Unknown parent groups (UPG) and metafounders (MF) were developed to address missing pedigrees in a population. This study aimed to incorporate MF into ssGBLUP models to select the best parents for controlled mating and the best progenies for cloning in a tree breeding population of Eucalyptus globulus. Methods: Genetic groups were defined to include base individuals of similar genetic origin. Tree growth was measured as total height (TH) and diameter at breast height (DBH), while disease resistance was assessed through heteroblasty (the transition from juvenile to adult foliage: ADFO). All traits were evaluated at 14 and 21 months. Two genomic multi-trait threshold linear models were fitted, with and without MF. Also, two multi-trait threshold-linear models based on phenotypic and pedigree information (ABLUP) were used to evaluate the increase in accuracy when adding genomic information to the model. To test the quality of models by cross-validation, the linear regression method (LR) was used. Results: The LR statistics indicated that the ssGBLUP models without MF performed better, as the inclusion of MF increased the bias of predictions. The ssGBLUP accuracy for both validations ranged from 0.42 to 0.68. Conclusions: The best model to select parents for controlled matings and individuals for cloning is ssGBLUP without MF.

Eucalyptus globulus Labill. is one of the most important species in the paper industry. Teratosphaeria nubilosa has affected plantations worldwide, infecting young foliage. Genome-wide association studies (GWASs) are essential to identify genomic segments associated with susceptibility to this disease. The inclusion of genomic strategies in breeding programs is key to the sustainability of the species. The aim of this study was to identify genomic regions associated with growth and heteroblasty (change from juvenile to adult foliage: ADFO) in a tree breeding population of E. globulus. Tree growth was measured as total height (TH) and diameter at breast height (DBH). All traits were evaluated at 14 and 21 months. A multi-trait threshold linear model was developed following the single-step genomic selection methodology. Genetic correlations (r_g) and narrow-sense heritability (h²) for all traits were estimated. Windows of 0.2 Mb were used. Only the windows with an estimated variance greater than 1% were considered. The r_g ranged from 0.51 to 0.97. The h² was high for ADFO (0.83–0.84) and lower for HT (0.37) and DBH (0.53). In growth traits, no QTLs were found that explained more than 1% of the variance. However, two genomic regions related to ADFO were identified on chromosomes 3 and 11.

Genomic selection is becoming an efficient and effective tool to increase genetic gain per unit time in forest tree breeding. The single-step genomic best linear unbiased prediction (ssGBLUP) method has been widely used in forest trees because it can use the complete phenotypic and pedigree data, with the genotypic information of a subset of the population. However, ssGBLUP assumes that all single nucleotide polymorphisms (SNP) explain the same amount of genetic variance, which is unlikely for traits whose major genes are segregating. The weighted ssGBLUP (WssGBLUP) method allows the use of unequal variance or weights for SNP in the construction of the genomic relationship matrix. The purpose of this study was to improve the accuracy of ssGBLUP by using the WssGBLUP, which gives more weight to relevant markers, for growth (tree height and diameter at breast height) and disease resistance traits (defoliation, leaf spot severity, and change to adult foliage) in a Eucalyptus globulus population. The weights were obtained iteratively by a nonlinear method that resembles Bayes A (nonlinear A). This weighting procedure was evaluated and compared with the unweighted ssGBLUP. The data comprised 3,853 individuals from 194 open-pollinated families, 975 of them were genotyped using the EUChip60K SNP chip. A random fold validation strategy was performed to evaluate the predictive ability of each model. Thus, a random sample of 10% of the genotyped trees (partial set) was removed and they were predicted based on the remaining data. This procedure was repeated 100 times. The predicted ability was estimated as the ratio of accuracies, defined as the Pearson correlation between the average prediction of the complete data and the partial set. For all traits, the ratio of accuracies of the weighting method was higher than that of the unweighting procedure. The average ratio between accuracies using ssGBLUP varied from 0.52 to 0.73 and 0.69 to 0.85 for growth and disease-related traits, respectively. These ratios were further increased from 0.59 to 0.76 for growth traits, and from 0.73 to 0.87 for disease related-traits by using the nonlinear A method under the WssGBLUP model. Our findings indicate that WssGBLUP has the potential to increase predictive abilities in genomic evaluations of Eucalyptus breeding programs.

Forest genetic breeding programs are intricate, long-term, and costly processes. To overcome these limitations, genomic selection emerges as an important tool to enable shorter cycles through early selection and to increase breeding value accuracies. Genomic selection is based on a predictive model derived from a training population (TP), integrating genotypic, phenotypic, and pedigree data to predict breeding values on selection candidates (SP) based on genomic information. As breeding generations advance, the relatedness between the TP and SP decreases, with the reduction of the prediction ability. The objective of this study is to optimize the TP for a Eucalyptus grandis breeding population across four consecutive generations (G1, G2, G3, and G4). A population of 66,221 individuals (from four generations) resulting from cross-breeding and open-pollination were measured for diameter at breast height (DBH) and total tree height (HT). Wood quality traits were determined for 16,068 trees and assessed using Near-Infrared Spectroscopy (NIRS). A total of 3,221 individuals from the four generations were genotyped using the EUCHIP60K or the Euc72k Axiom Eucalyptus Genotyping Array. We compared the prediction ability using all the previous generations in the TP as opposed to only including the previous generation. Single-step GBLUP (ssGBLUP) models were evaluated in three phenotypic traits: volume, pulp yield, and wood density. The inclusion of all the previous generations provides a higher predictive ability of 0.40 and 0.54 for volume and pulp yield, respectively, in the target generation G4. When only the previous generation G3 was used as a TP, no significant differences in predictive abilities were observed for volume, and there was a slight decrease in pulp yield. In contrast, for wood density, the predictive ability was highest (0.37) in the G4 when the model was training with only the G3, while including all preceding generations decreased to 0.33. Our results provide insight into the optimization of TP across generations, encouraging the implementation of genomic selection in the Eucalyptus breeding program.