Research highlights, The Plant Genome, Volume 16, Issue 1

Drought stress (DS) adversely affects agricultural productivity worldwide and is expected to rise in the coming years. Therefore, it is vital to understand the physiological, biochemical, and molecular mechanisms associated with DS. Raza et al. (https://doi.org/10.1002/tpg2.20279) examined recent advances in plant responses to DS to expand our understanding of DS-associated mechanisms. To cope with DS, we recommend incorporating several approaches, such as multi-omics (genomics, transcriptomics, proteomics, metabolomics, and epigenomics), genome editing, conventional and speed breeding, phytohormone treatments, and agronomic platforms, to develop drought-smart cultivars to achieve the ‘zero hunger’ goal.


COMPARING ARTIFICIAL INTELLIGENCE TECHNIQUES WITH THE STATE-OF-THE-ART PARAMETRIC PREDICTION MODELS FOR PREDICTING SOYBEAN TRAITS
Soybean is one of the major crops in agriculture, and one of the major tasks in soybean breeding is to develop superior lines in terms of yield, protein, and oil content. Genomic prediction is a technique where genotypic and phenotypic information are used to predict the performance of unobserved lines in fields to aid the selection process. In this study, Ray et al. (https://doi.org/10.1002/tpg2.20263) compared the state-of-the-art parametric genomic prediction models with nonparametric artificial intelligence-based techniques in terms of prediction accuracy. Here, prediction accuracy is defined as the correlation between the true (or observed) and the predicted phenotypic values within the same environment. They concluded that in general the conventional model performed the best when the genotypic and environmental main effects, and the genotype-by-environment interaction were formally included into the model. Moreover, their conclusions are agreed with other similar studies showing that the performance of the artificial intelligence-based models is highly influenced by the tuning of the hyperparameters and the size of the training data; thus, further investigation is necessary.

UNDERSTANDING FLOWERING TIME BETTER USING NOVEL PHENOTYPES
Improved understanding of the genetics underlies when plants' flower is key to broadening genetic diversity and overcomes adaptation constraints in plant-breeding programs. Neupane et al. (https://doi.org/10.1002/tpg2.20269) took flowering time and environmental data from a trial of diverse lentil germplasm grown in multiple locations around the world to reveal genes associated with this trait. To do this, they derived latent variable phenotypes from mathematical models for use as main traits and covariates. Regions of the genome associated with response to temperature and photoperiod were identified that reinforced the role of known genes and pointed to previously unknown associations. Their approach can be replicated with other crop species for understanding complex traits such as flowering time.

TARGETED GENOTYPING DEVELOPED FOR SOYBEANS
Enabling genomic selection for soybean progeny rows requires a genome-wide genotyping method that provides a low per-sample cost. Molecular inversion probe (MIP) is a targeted genotyping method that is low cost and can provide high-quality data. Wang et al. (https://doi.org/10. 1002/tpg2.20270) reported developing a 1K MIP SNP set with markers selected to maximize genetic polymorphism when used to screen soybean-breeding populations. The MIP set was found to provide accurate genotyping and coverage across the genome in elite soybean populations. MIPs will be a powerful tool that will enable genomic selection within soybean-breeding programs.

6
Phytophthora RESISTANCE IN DIVERSE STRAWBERRIES Phytophthora cactorum causes crown rot, a devastating disease of global importance in cultivated strawberry. Jiménez et al. (https://doi.org/10.1002/tpg2.20275) reported a single large-effect locus, RPc2, contributing to genetic resistance to Phytophthora crown rot in strawberry. RPc2 alone accounts for up to 50% of the genetic variance and improves genomic prediction model accuracy by up to 15% when included as a fixed effect in diverse strawberry germplasm. Although complete genetic resistance is rare, the beneficial allele of RPc2 is present in all modern University of California-Davis germplasm. Thus, marker-assisted and genomic selection incorporating RPc2 may be robust solutions for improving and identifying donors of genetic resistance to Phytophthora crown rot in strawberry.

LIGNIN INTERWOVEN IN MAIZE DEFENSE
Lignin plays an integral part in plant structure, function, and defense. In maize, brown midrib (bm) mutants define a desirable market class of silage based on lower lignin levels that result increased digestibility in ruminant livestock. Kolkman et al. (https://doi.org/10.1002/tpg2.20278) showed that bm mutants (bm1-bm4) are also more susceptible to foliar fungal (northern leaf blight [NLB], gray leaf spot, and anthracnose leaf blight) and bacterial (Stewart's wilt) diseases, and in some cases susceptible to Gibberella ear rot and anthracnose stalk rot. We examine the previously described quantitative trait loci (QTLs) and correlations between disease traits and stalk strength, as a proxy for lignin, to determine if there is evidence for pleiotropy. More specifically, a high resolution genome-wide association study for resistance to NLB revealed candidate genes for resistance to NLB with diverse functions, including lignin-and nonlignin-related genes. Lignin-related genes ranged from developmental genes, transcription factors, monolignol, and monolignol processing genes. Breeding for resistance in a market class defined by a biochemical pathway, such as low lignin silage maize, can be achieved through selection of resistance genes from nontarget pathways.

GENOMIC SELECTION IN TEA BREEDING
Tea-breeding programs traditionally use phenotype information to identify the best individuals as parents of next generations or as potential new varieties evaluated in extensive field trials. This phenotype-based selection has a limited accuracy and is also a time-consuming-it takes about 16 years to develop a new variety for commercial release. Genomic selection first trains a prediction model on a population of genotyped and phenotyped individuals and then uses that model to predict breeding values of selection candidates early in a breeding cycle using only genome data. Lubanga et al. (https://doi.org/10.1002/tpg2.20282) showed how to leverage genomic selection in tea breeding with two competitive approaches. When applied at the early seedling stage, genomic selection is predicted to generate 1.7 times more genetic gain than the traditional phenotype-based selection using the same budget.

CARBON 13 RATIO PLASTICITY IDENTIFIED IN SOYBEAN
Water use efficiency (WUE) is potentially an important trait for improving soybean drought tolerance, but WUE is difficult to determine in field experiments and increased WUE may decrease growth and yield when soil moisture is not limiting. The ability to change WUE, or C13 ratio, in response to the environment is called plasticity and could potentially overcome the problem of low yields in favorable environments for genotypes with high WUE. Chamarthi et al. (https://doi.org/ 10.1002/tpg2.20284) used the ratio of C13 to C12 (C13 ratio) in plant tissue as a surrogate measure of WUE and evaluated 473 accessions for C13 ratio in several diverse environments. Chamarthi et al. identified accessions that were highly plastic for C13 ratio and used association mapping to identify chromosomal regions closely associated with C13 ratio and with plasticity. These results represent a genetic resource for improving soybean drought tolerance.

CHARACTERIZATION OF Rsg3, A NOVEL GREENBUG RESISTANCE GENE FROM THE CHINESE BARLEY LANDRACE PI 565676
Greenbug (Schizaphis graminum Rondani) is a pest that poses a serious threat to cereal production worldwide, and resistance genes are urgently needed to enhance greenbug resistance in barley. Xu et al. (https://doi.org/10.1002/tpg2.20287) discovered a novel greenbug resistance gene, Rsg3, in the Chinese landrace PI 565676 and mapped Rsg3 to an interval of <100 kb, in which four genes were annotated. An allelism test indicated that Rsg3 is independent of the Rsg1 locus, with an estimated recombination frequency of 12.85% ± 0.20% and a genetic distance of 13.14 ± 0.21 cM between the two loci. Two SNPs flanking Rsg3 were converted to Kompetitive Allele Specific PCR markers, which can be used to tag Rsg3 in barley breeding.

META-QTL ANALYSIS FOR CHLOROPHYLL TRAITS OF WHEAT
Guo et al. (https://doi.org/10.1002/tpg2.20294) identified a total of 56 meta-QTLs (MQTLs) for chlorophyll traits by QTL meta-analysis, more than half of which were validated by marker-trait associations in genome-wide associations studies. Six MQTLs were selected as breeders' MQTLs. In addition, they have identified 21 candidate genes that are significantly enriched in chlorophyll metabolism and photosynthetic pathways, mainly related to chlorophyll metabolizing enzymes and photosynthetic proteins.

SCREENING OF HD-ZIP GENES RELATED TO PRICKLE DEVELOPMENT IN Zanthoxylum armatum
Plant prickles can play a protective role. On the other hand, they cause a lot of trouble in production. Zhang et al. (https:// doi.org/10.1002/tpg2.20295) identified 76 homeodomain leucine zipper (HD-ZIP) genes from the genome of Zanthoxylum armatum and divided them into four subfamilies (I-IV) based on phylogenetic analysis, conservative motif composition, and gene structure analysis. In addition, the chromosome location, synteny, and cis-elements of ZaHDZ gene were analyzed. Among all ZaHDZ genes, the ZaHDZ16 gene was specifically highly expressed in the prickles and was found to be highly homologous to some HD-ZIP genes related to