Molecular Genetics Enhances Plant Breeding

Human-driven plant selection, a practice as ancient as agriculture itself, has laid the foundations of plant breeding and contemporary farming [...].

Human-driven plant selection, a practice as ancient as agriculture itself, has laid the foundations of plant breeding and contemporary farming [1]. The principles of classical breeding still comprise the nucleus of modern crop science and agricultural production [2]. Recent unthinkable methodological achievements in molecular screening, genome sequencing, 'omic' technologies, trans-genesis, and computational power have advanced disciplinary boundaries in order to revolutionize the field of varietal development and its broader connections with related disciplines. These advances have opened up new trans-disciplinary arenas in which classical plant improvement intersects with genomics, molecular biology, biotechnology, and bioinformatics [3]. Yet, the factual potential of these interplays is often disregarded by prevailing atomized views and skepticism, both of which penalize the dawn of emerging properties and their ultimate deployment into framers' fields [4]. Therefore, this second Special Issue of IJMS on "Molecular Genetics and Plant Breeding" exemplifies through 31 inter-disciplinary works (https://www.mdpi.com/journal/ijms/special_issues/gene_plant_breading_2nd, accessed on 6 June 2023) the ways in which disruptive molecular research opens up innovative pathways for the exploration, transformation, and utilization of crop biodiversity in order to improve essential breeding targets and replenish the genetic potential of the cultivated genepools in order to achieve greater yield and sustainability [5].

Marker-Guided Pre-Breeding
Natural genetic variation continues to constitute a major source of crop innovation, and its exploration remains a key milestone in molecular pre-breeding. This Special Issue provides some insights in this regard by presenting a series of molecular characterizations of diverse germplasm (Table 1). For instance, López-Hernández and Cortés [6] assessed the scale and determinants of somaclonal coding diversity in mint (Mentha spp.), introduced to the northern Andes, using RNA-seq on 29 clonally propagated accessions. The authors found a single genetic cluster for M. × piperita, and three clusters for M. spicata, suggesting two independent introductions of the latter.
Meanwhile, marker-assisted pre-breeding is underdoing a transformation from the method of classically screening crop genepools to the assessment of their associated antagonistic biotic agents. For example, Choi et al. [7] elucidated the molecular phylogenetic origin of seven strains of soybean mosaic virus (SMV) that had been sampled from 150 Glycine max L. accessions. The use of RNA-seq and 143 SMV available genomes enabled concluding that recombination and plant hosts drive the genetic diversity of SMV.   Similarly, Jo et al. [12] systematically examined the viromes of 15 pepper (Capsicum annuum L.) cultivars through RNA-seq, enabling a high-throughput identification of the principal viromes present in commercially important pepper genotypes. As a promising perspective, these initial characterizations of natural variation may unleash a further potential when coupled with explicit estimates of their genomic bases, i.e., the genetic determinants of trait variation.
As such, genetic mapping arises as a key step in the pre-breeding pipeline that enables the genetic architecture of key traits to be disclosed, and ultimately provides candidate markers and genes for further marker-guided applications, such as parental screening, marker-assisted selection [37], and gene editing [38]. In this regard, and as part of this Special Issue, Kim et al. [13] genotyped 120 double haploid rice lines with 778 SSRs in order to reconstruct the genetic basis of germination under cold conditions. The authors managed to retrieve 4 QTLs and 41 genes, 25 of which were validated via qRT-PCR. Similarly, Dossou et al. [14] characterized 410 sesame (S. indicum) accessions using WGR, and performed GWAS for lignan-lignin biosynthesis. The team found that SiNST1 is a major target gene for the molecular breeding of lignan content. In soybean, Ouyang et al. [15] and Seo et al. [16] performed WGR of RIL families with ca. 155 genotypes each. The teams, respectively, identified five QTLs for drought tolerance, and validated accuracies above 90% for existing QTLs for pod-shattering tolerance. Finally, Mokrzycka et al. [17] and Zhou et al. [18], respectively, genotyped 509 wheat accessions and 300 RILs with 13,499 DArT-SNPs and a 55 K SNP array. The authors discovered 1261 candidate markers for six agronomical traits, and the QSlu.sicau-2B-2 MAS target for spike layer uniformity. A common denominator across all genetic mapping exercises is the need for a further validation of the associated markers [39] before any downstream plant improvement application. After all, they are prone to displaying inflated rates of false positives due to multiple comparison testing, population stratification, and intrinsic redundancy from linkage disequilibrium (LD) [40].
Marker validation often requires that a battery of downstream analyses be performed. The first step required is narrow mapping of the causal variants using target genotyping across extended panels. A second, more experimental approach involves examining the expression profiles of flanking candidate genes in terms of congruency and stability. Several teams pursued this goal as part of the current Special Issue. For example, Yang et al. [19] and Zhang et al. [20] used the Columbia-0 (Col) A. thaliana ecotype to validate the functional roles of PIF4 and AtEAU (1 and 2) genes, respectively, through transcriptomic tools as boosters and repressors of basal thermotolerance and ABA response. Meanwhile, Kong et al. [21] and Zhang et al. [22], respectively, implemented epigenomics and qRT-PCR in cabbage (B. oleracea) and Chinese cabbage (B. rapa) to study the DNA methylation footprint during microspore embryogenesis induced by heat shock, and the functional role of BrGOLDEN in carotenoid biosynthesis. Interestingly, rapeseed (B. napus) was a widely considered study system by Du et al. [24], Wahid et al. [25], and Zhang et al. [26], who used qRT-PCR to corroborate the functionality of BnaZSNRT2s, BnARI, and BnaTH genes for nitrogen uptake, agronomical trait regulation, and abiotic stress tolerance, respectively. Other studies focused on key crops for food security. For example, while also considering the ABA pathway, Yu et al. [23] used qRT-PCR to demonstrate the up-regulation of PHO-encoding genes by ABA in maize and rice. On the other hand, Xiao et al. [29] validated SbDOF21 as a key regulator for starch biosynthesis in sorghum, and Guo et al. [30] characterized a new SPR-repressor Q allele (Q c5 ) for compact spikes and bread quality in wheat. The other studies which stand out are those by Mažeikienė et al. [27] and Liu et al. [28], who brought RNA-seq technology to the new promissory crops Ribes nigrum and Sacha Inchi (Plukenetia volubilis). Each team was able to provide transcriptomic resources to breed for tolerance to the BRV virus [27] and better comprehend the expression profiles of seeds at two different developmental stages [28].
Finally, the works by Hu et al. [31] and Huang et al. [32] utilized trans-genesis to move beyond the species boundary and validate in an Arabidopsis background candidate genes for cold (HuNAC20 and 25) and osmatic (MfWRKY40) stress tolerance, respectively, from two other exotic crops, namely, the cactus-fruit Pitaya (Hylocereus monacanthus) [31] and the woody resurrection plant (Myrothamnus flabellifolia) [32]. Despite these advances in germplasm characterization, genetic mapping, and functional validation, moving forward, elite cultivars are beginning to benefit from embracing genomic-enabled breeding.

Molecularly Enabled Breeding
Mobilizing germplasm's allelic novelty, leveraging marker genetic mapping, and functionally corroborating candidate genes to customize crops requires the modernization of classical plant breeding with phenomic and genomic tools [2]. Perhaps one of the promptest strategies for bridging long breeding cycles is trans-genesis. When it targets standing variation, segregating within the same species genepool, it simply provides a shortcut to the time-consuming alternative of recurrent backcrossing [37]. However, it also allows customizability for de novo variation. This Special Issue compiles an exquisite repertoire in this regard. For instance, Guo et al. [33] illustrated, once again in rapeseed, that it is feasible to genetically engineer BnaC02.TMT.a, a γ-TMT paralogue, for salt tolerance breeding via the use of homozygous transgenic lines. Similarly, Liu et al. [34] silenced GmBIR1 with BPMV-VIGS to demonstrate that it is a negative regulator of immunity in soybean. Finally, Chen et al. [35] went one step further and explored mutagenesis as a reliable source of de novo allelic variation. The authors characterized mutant Q alleles (Q s1 and Q c1 -N8) for grain yield and grain protein in the wheat cultivar 'Shumai482' and its S-Cp1-1 mutant [35], complementing the results achieved by Guo et al. [30] in terms of compact spikes and bread quality. In spite of these exciting results achieved for the development of trans-genesis, classical intercrossing has not lost validity as a breeding strategy, but rather it has been permeable to the latest genomic advances, such as marker-assisted backcrossing (MABC) [37], introgression breeding, and genomic-assisted recurrent selection [41].
Inter-specific introgression breeding, in particular, aims to break species boundaries to pyramid exotic variation from one species to elite commercial varieties from the other species [37]. However, inter-specific crossing conveys two main challenges, which are species incompatibility and polygenic trait variation. Therefore, coupling hybrid breeding with bridge genotypes and guiding molecular markers unleashes novel opportunities to improve the chances of success, pace, and precision of the target introgression. Such utility has been reinforced in rice by Zhang et al. [36], who bred for heat tolerance by introgressed sorghum ancestry. The authors utilized the rice restorer line 'R21' and the recipient restorer line Jin 'Hui 1', and were able to demonstrate with WGR that the sorghum introgression in line 'R21' confers mobile heat tolerance. Gene editing may speed up rice breeding for abiotic stresses [42], too. Bean breeders are also seeing a similar development by GBS-genotyping 87 advanced lines with inter-specific ancestries between common bean (P. vulgaris) and tepary bean (P. acutifolius) for heat and drought tolerance across four environments in coastal northern South America [43]. The authors found 47 associated loci and 90 flanking candidate genes for molecular-guided downstream selection. Meanwhile, the team also detected suitable allelic variation within the candidate genes of tepary bean (P. acutifolius) that transcends the adaptive genepool of common bean (P. vulgaris) [44]. These studies demonstrate that the integration of genomic-and gene-based strategies can leverage inter-specific adaptive variation via bridge genotypes in order to deliver candidate introgressed lines for heat tolerance [43]. In certain cases, grafting could also provide a fast track to harness such species diversity [45]. These success stories that intermingle modern genotyping technology with classical intercrossing exemplify, against any intellectual skepticism, the factual applicability of the molecular breeding paradigm. Still, there remains room for further developments.

Perspectives
As molecular and phenomic data continue to pile up, modern analytical techniques must be embraced. Predictive breeding, reached throughout genomic-enabled selection [41] and machine learning [11] at the interface of the plant breeding triangle [46], confers a primary way forward capable of bringing together molecular biology and plant genetic paradigms [47]. In this regard, Khan et al. [11] reviewed applications of crop breeding in this Special Issue, calling for a better integration of AI with "omics" tools. Still, effective crop mobilization networks and open access data must be assured in order to build sufficiently reliable training datasets without sampling bias or over-fitting prediction [48].
Meanwhile, biotic and abiotic stresses are becoming more common, jeopardizing global food production. In this sense, another review within this collection by Zhang et al. [8] was more concrete in envisioning gene targets for biotic pressures, specifically that clubroot control methods in cruciferous crops could be harnessed with allelic variation at the R genes, which are in turn susceptible to be gene-edited, eventually. Similarly, the reviews by Chen et al. [9] and Jiang et al. [10] prospected gene-enabled abiotic stress responses by looking at the MYB genes in plant roots and the miR393 during plant development, respectively.
The method of screening the genetic determinants of the abiotic susceptibility, and its potential gene editing, should be supplemented by explicit in situ eco-physiological indices targeting specific stresses, performing potential ecological niche modeling (ENM) under present and future scenarios as part of climate vulnerability assessments, undertaking gap analyses of the available variation for pre-breeding, exploring genomic selection signatures of historic adaptation, and unpicking the genome-environment associations (GEA) of current niche preferences [3]. After all, tackling the climate crisis and agrobiodiversity loss in the process of addressing food security [5] demands that we harness trans-disciplinary sustainable enterprises that promote integrative agendas among the otherwise disentangled fields of in situ and ex situ conversation, physiology, ecology, molecular genetics, plant breeding [2], conservation, seed delivery [4], food policy [49], and marketing.
Author Contributions: A.J.C. and H.D. conceived this Special Issue and jointly invited potential authors, handled manuscripts, recommended reviewers, and approved submissions. A.J.C. prepared a first draft, which was later on edited by H.D. All authors have read and agreed to the published version of the manuscript.