Significance: Consumers often regard heirloom fruit varieties grown in the garden as more flavorful than commercial varieties purchased at the grocery store. While plant breeders have historically focused on improving producer-orientated traits such as yield, consumer-oriented traits such as flavor have regularly been neglected. This is, in part, due to the difficulty associated with measuring the sensory perceptions of flavor. Here, we combine fruit chemical and consumer sensory panel information to train machine learning models that can predict how flavorful a fruit will be from its chemistry. By increasing the throughput of flavor evaluations, these models will help plant breeders to integrate flavor earlier in the breeding pipeline and aid in the design of varieties with exceptional flavor profiles.

Breeding crops for improved flavor is challenging due to the high cost of sensory evaluation and the difficulty of connecting sensory experience to chemical composition. The main goal of this study was to identify the chemical drivers of sweetness and consumer liking for fresh strawberries (Fragaria × ananassa). Fruit of 148 strawberry samples from cultivars and breeding selections were grown and harvested over seven years and were subjected to both sensory and chemical analyses. Each panel consisted of at least 100 consumers, resulting in more than 15,000 sensory data points per descriptor. Three sugars, two acids and 113 volatile compounds were quantified. Consumer liking was highly associated with sweetness intensity, texture liking, and flavor intensity, but not sourness intensity. Partial least square analyses revealed 20 volatile compounds that increased sweetness perception independently of sugars; 18 volatiles that increased liking independently of sugars; and 15 volatile compounds that had positive effects on both. Machine learning-based predictive models including sugars, acids, and volatiles explained at least 25% more variation in sweetness and liking than models accounting for sugars and acids only. Volatile compounds such as γ-dodecalactone; 5-hepten-2-one, 6-methyl; and multiple medium-chain fatty acid esters may serve as targets for breeding or quality control attributes for strawberry products. A genetic association study identified two loci controlling ester production, both on linkage group 6 A. Co-segregating makers in these regions can be used for increasing multiple esters simultaneously. This study demonstrates a paradigm for improvement of fruit sweetness and flavor in which consumers drive the identification of the most important chemical targets, which in turn drives the discovery of genetic targets for marker-assisted breeding.

Sweet corn is one of the most important vegetables in the United States and Canada. Here, we present a de novo assembly of a sweet corn inbred line Ia453 with the mutated shrunken2-reference allele (Ia453-sh2). This mutation accumulates more sugar and is present in most commercial hybrids developed for the processing and fresh markets. The ten pseudochromosomes cover 92% of the total assembly and 99% of the estimated genome size, with a scaffold N50 of 222.2 Mb. This reference genome completely assembles the large structural variation that created the mutant sh2-R allele. Furthermore, comparative genomics analysis with six field corn genomes highlights differences in single-nucleotide polymorphisms, structural variations, and transposon composition. Phylogenetic analysis of 5,381 diverse maize and teosinte accessions reveals genetic relationships between sweet corn and other types of maize. Our results show evidence for a common origin in northern Mexico for modern sweet corn in the U.S. Finally, population genomic analysis identifies regions of the genome under selection and candidate genes associated with sweet corn traits, such as early flowering, endosperm composition, plant and tassel architecture, and kernel row number. Our study provides a high-quality reference-genome sequence to facilitate comparative genomics, functional studies, and genomic-assisted breeding for sweet corn.

Estimation of allele dosage, using genomic data, in autopolyploids is challenging and current methods often result in the misclassification of genotypes. Some progress has been made when using SNP arrays, but the major challenge is when using next generation sequencing data. Here we compare the use of read depth as continuous parameterization with ploidy parameterizations in the context of genomic selection (GS). Additionally, different sources of information to build relationship matrices were compared. A real breeding population of the autotetraploid species blueberry (Vaccinium corybosum), composed of 1,847 individuals was phenotyped for eight yield and fruit quality traits over two years. Continuous genotypic based models performed as well as the best models. This approach also reduces the computational time and avoids problems associated with misclassification of genotypic classes when assigning dosage in polyploid species. This approach could be very valuable for species with higher ploidy levels or for emerging crops where ploidy is not well understood. To our knowledge, this work constitutes the first study of genomic selection in blueberry. Accuracies are encouraging for application of GS for blueberry breeding. GS could reduce the time for cultivar release by three years, increasing the genetic gain per cycle by 86% on average when compared to phenotypic selection, and 32% when compared with pedigree-based selection. Finally, the genotypic and phenotypic data used in this study are made available for comparative analysis of dosage calling and genomic selection prediction models in the context of autopolyploids.

Alfalfa is planted in more than 30 million hectares worldwide, but despite its popularity in temperate regions, it is not widely grown in subtropical agroecosystems. It is critical to improve alfalfa for such regions, considering current predictions of global warming and the increasing demands for animal-based products. In this study, we examined the diversity present in subtropical alfalfa germplasm and reported genetic parameters for forage production. An initial screening was performed from 2014 to 2016, evaluating 121 populations from different subtropical origins. Then, a breeding population was created by crossing selected plants, resulting in 145 full-sib and 36 half-sib families, which were planted in a row-column design with augmented representation of three controls (‘Bulldog805′, ‘FL99′ and ‘UF2015′). Dry matter yield (DMY), canopy height (AH), and percentage blooming (BLOOM) were measured across several harvests. Moderate narrow-sense heritability and high genetic correlations between consecutive harvests were estimated for all traits. The breeding line UF2015 produced higher DMY than FL99 and Bulldog805, and it could be a candidate cultivar release. Several families produced higher DMY than all checks, and they can be utilized to develop high yielding and adapted alfalfa cultivars for subtropical agroecosystems.