Multi-modal deep learning improves grain yield prediction in wheat breeding by fusing genomics and phenomics.

MOTIVATION: Developing new crop varieties with superior performance is highly important to ensure robust and sustainable global food security. The speed of variety development is limited by long field cycles and advanced generation selections in plant breeding programs. While methods to predict yield from genotype or phenotype data have been proposed, improved performance and integrated models are needed. RESULTS: We propose a machine learning model that leverages both genotype and phenotype measurements by fusing genetic variants with multiple data sources collected by unmanned aerial systems. We use a deep multiple instance learning framework with an attention mechanism that sheds light on the importance given to each input during prediction, enhancing interpretability. Our model reaches 0.754 ± 0.024 Pearson correlation coefficient when predicting yield in similar environmental conditions; a 34.8% improvement over the genotype-only linear baseline (0.559 ± 0.050). We further predict yield on new lines in an unseen environment using only genotypes, obtaining a prediction accuracy of 0.386 ± 0.010, a 13.5% improvement over the linear baseline. Our multi-modal deep learning architecture efficiently accounts for plant health and environment, distilling the genetic contribution and providing excellent predictions. Yield prediction algorithms leveraging phenotypic observations during training therefore promise to improve breeding programs, ultimately speeding up delivery of improved varieties. AVAILABILITY AND IMPLEMENTATION: Available at https://github.com/BorgwardtLab/PheGeMIL (code) and https://doi.org/doi:10.5061/dryad.kprr4xh5p (data).

Wheat genetic resources have avoided disease pandemics, improved food security, and reduced environmental footprints: A review of historical impacts and future opportunities.

The use of plant genetic resources (PGR)-wild relatives, landraces, and isolated breeding gene pools-has had substantial impacts on wheat breeding for resistance to biotic and abiotic stresses, while increasing nutritional value, end-use quality, and grain yield. In the Global South, post-Green Revolution genetic yield gains are generally achieved with minimal additional inputs. As a result, production has increased, and millions of hectares of natural ecosystems have been spared. Without PGR-derived disease resistance, fungicide use would have easily doubled, massively increasing selection pressure for fungicide resistance. It is estimated that in wheat, a billion liters of fungicide application have been avoided just since 2000. This review presents examples of successful use of PGR including the relentless battle against wheat rust epidemics/pandemics, defending against diseases that jump species barriers like blast, biofortification giving nutrient-dense varieties and the use of novel genetic variation for improving polygenic traits like climate resilience. Crop breeding genepools urgently need to be diversified to increase yields across a range of environments (>200 Mha globally), under less predictable weather and biotic stress pressure, while increasing input use efficiency. Given that the ~0.8 m PGR in wheat collections worldwide are relatively untapped and massive impacts of the tiny fraction studied, larger scale screenings and introgression promise solutions to emerging challenges, facilitated by advanced phenomic and genomic tools. The first translocations in wheat to modify rhizosphere microbiome interaction (reducing biological nitrification, reducing greenhouse gases, and increasing nitrogen use efficiency) is a landmark proof of concept. Phenomics and next-generation sequencing have already elucidated exotic haplotypes associated with biotic and complex abiotic traits now mainstreamed in breeding. Big data from decades of global yield trials can elucidate the benefits of PGR across environments. This kind of impact cannot be achieved without widescale sharing of germplasm and other breeding technologies through networks and public-private partnerships in a pre-competitive space.

High-density mapping of durable and broad-spectrum stripe rust resistance gene Yr30 in wheat.

KEY MESSAGE: The durable stripe rust resistance gene Yr30 was fine-mapped to a 610-kb region in which five candidate genes were identified by expression analysis and sequence polymorphisms. The emergence of genetically diverse and more aggressive races of Puccinia striiformis f. sp. tritici (Pst) in the past twenty years has resulted in global stripe rust outbreaks and the rapid breakdown of resistance genes. Yr30 is an adult plant resistance (APR) gene with broad-spectrum effectiveness and its durability. Here, we fine-mapped the YR30 locus to a 0.52-cM interval using 1629 individuals derived from residual heterozygous F5:6 plants in a Yaco"S"/Mingxian169 recombinant inbred line population. This interval corresponded to a 610-kb region in the International Wheat Genome Sequencing Consortium (IWGSC) RefSeq version 2.1 on chromosome arm 3BS harboring 30 high-confidence genes. Five genes were identified as candidate genes based on functional annotation, expression analysis by RNA-seq and sequence polymorphisms between cultivars with and without Yr30 based on resequencing. Haplotype analysis of the target region identified six haplotypes (YR30_h1-YR30_h6) in a panel of 1215 wheat accessions based on the 660K feature genotyping array. Lines with YR30_h6 displayed more resistance to stripe rust than the other five haplotypes. Near-isogenic lines (NILs) with Yr30 showed a 32.94% higher grain yield than susceptible counterparts when grown in a stripe rust nursery, whereas there was no difference in grain yield under rust-free conditions. These results lay a foundation for map-based cloning Yr30.

Polar-Effect-Directed Control in Site-Selectivity of Radical Substitution Enables C-H Perfluoroalkylation of Coumarins.

A novel Ru-catalyzed protocol for C-7 selective C-H trifluoromethylation of coumarins in the presence of light is presented. This reaction undergoes a radical type nucleophilic substitution instead of a radical type electrophilic substitution owing to the benzocore activation as a result of lowering the lowest unoccupied molecular orbital (LUMO).

An Expedient Approach to Access Phenalenones via Unconventional [1,2]-Phospha-Brook Rearrangement/Carbonyl Migration.

An efficient protocol for the synthesis of 2,3-disubstituted phenalenones from para-quinone methides (p-QMs) and acenaphthoquinone is described. The reaction involves P(NMe2)3-mediated [1,2]-phospha-Brook rearrangement followed by Lewis acid-assisted 1,2-carbonyl migration to afford the 2,3-disubstituted phenalenones. The developed protocol tolerates a broad range of substrates to form a variety of phenalenones in good to excellent yields. Moreover, the utility of the synthesized phenalenones is also demonstrated by performing its transformations to other adducts.

Deciphering Asymmetric Brønsted Base-Aminocatalytic Mode in Pudovik/[1,2]-Phospha-Brook Rearrangement/Michael Cascade Reaction.

An approach involving the use of a bifunctional aminocatalyst containing Brønsted base and iminium activation sites for asymmetric multicomponent reactions involving [1,2]-phospha-Brook rearrangement has yet to be realized. Herein, we present an aminocatalytic enantioselective conjugate addition of α-phosphonyloxy enolates formed via [1,2]-phospha-Brook rearrangement to α,ß-unsaturated ketones. The methodology unfolds a simple one-pot operation consisting of a robust additive-free catalytic system providing a series of oxindole derivatives having two contiguous stereocenters in high yields with excellent stereoselectivities.

Photoinduced metal-free trifluoro/perfluoroalkylation of heteroarenes.

A practical and straightforward protocol to access trifluoromethylated/perfluoroalkylated heteroarenes via radical-type nucleophilic substitution rather than typical radical-type electrophilic substitution is described here. The substrate scope was observed to be broad and diverse-covering arenes, heteroarenes (containing N, O, S), bioactive cores, and allylic cores. Mechanistic studies confirmed a radical-mediated reaction pathway.

Genetic Analysis of Stripe Rust Resistance in the Chinese Wheat Cultivar Luomai 163.

Stripe or yellow rust (YR) caused by Puccinia striiformis tritici (Pst) is an important foliar disease affecting wheat production globally. Resistant varieties are the most economically and environmentally effective way to manage this disease. The common winter wheat (Triticum aestivum L.) cultivar Luomai 163 exhibited resistance to Pst races CYR32 and CYR33 at the seedling stage and showed a high level adult plant resistance in the field. To understand the genetic basis of YR resistance in this cultivar, 142 F5 recombinant inbred lines (RILs) derived from cross Apav#1 × LM163 and both parents were genotyped with the 16K SNP array and bulked segregant analysis sequencing (BSA-Seq). The analysis detected a major gene, YrLM163, at the seedling stage associated with the 1BL.1RS translocation. Additionally, three genes for resistance at the adult plant stage were detected on chromosome arms 1BL (Lr46/Yr29/Pm39/Sr58), 6BS and 6BL in Luomai 163, whereas Apav#1 contributed resistance at a QTL on 2BL. These QTL explained YR disease severity variations ranging from 6.9 to 54.8%. KASP markers KASP-2BL, KASP-6BS and KASP-6BL for three novel loci QYr.hzau-2BL, QYr.hzau-6BS and QYr.hzau-6BL were developed and validated. QYr.hzau-1BL, QYr.hzau-2BL and QYr.hzau-6BS showed varying degrees of resistance to YR when present individually or in combination based on genotype and phenotype analysis of a panel of 570 wheat accessions. Six RILs combining resistance alleles of all QTL, showing higher resistance to YR in the field than Luomai 163 with disease severities of 10.7-16.0%, are important germplasm resources for breeding programs to develop YR resistant wheat varieties with good agronomic traits.

Genetic analysis of stripe rust resistance in the common wheat line Kfa/2*Kachu under a Chinese rust environment.

KEY MESSAGE: We mapped a new race-specific seedling stripe rust resistance gene on wheat chromosome 5BL and a new APR locus QYr.hazu-2BS from CIMMYT wheat line Kfa/2*Kachu. Breeding resistant wheat (Triticum aestivum) varieties is the most economical and efficient way to manage wheat stripe rust, but requires the prior identification of new resistance genes and development of associated molecular markers for marker-assisted selection. To map stripe rust resistance loci in wheat, we used a recombinant inbred line population generated by crossing the stripe rust-resistant parent 'Kfa/2*Kachu' and the susceptible parent 'Apav#1'. We employed genotyping-by-sequencing and bulked segregant RNA sequencing to map a new race-specific seedling stripe rust resistance gene, which we designated YrK, to wheat chromosome arm 5BL. TraesCS5B02G330700 encodes a receptor-like kinase and is a high-confidence candidate gene for YrK based on virus-induced gene silencing results and the significant induction of its expression 24 h after inoculation with wheat stripe rust. To assist breeding, we developed functional molecular markers based on the polymorphic single nucleotide polymorphisms in the coding sequence region of YrK. We also mapped four adult plant resistance (APR) loci to wheat chromosome arms 1BL, 2AS, 2BS and 4AL. Among these APR loci, we determined that QYr.hazu-1BL and QYr.hazu-2AS are allelic to the known pleiotropic resistance gene Lr46/Yr29/Pm39 and the race-specific gene Yr17, respectively. However, QYr.hazu-2BS is likely a new APR locus, for which we converted closely linked SNP polymorphisms into breeder-friendly Kompetitive allele-specific PCR (KASP) markers. In the present study, we provided new stripe rust resistance locus/gene and molecular markers for wheat breeder to develop rust-resistant wheat variety.

High density mapping of wheat stripe rust resistance gene QYrXN3517-1BL using QTL mapping, BSE-Seq and candidate gene analysis.

KEY MESSAGE: Fine mapping of a major stripe rust resistance locus QYrXN3517-1BL to a 336 kb region that includes 12 candidate genes. Utilization of genetic resistance is an effective strategy to control stripe rust disease in wheat. Cultivar XINONG-3517 (XN3517) has remained highly resistant to stripe rust since its release in 2008. To understand the genetic architecture of stripe rust resistance, Avocet S (AvS) × XN3517 F6 RIL population was assessed for stripe rust severity in five field environments. The parents and RILs were genotyped by using the GenoBaits Wheat 16 K Panel. Four stable QTL from XINONG-3517 were detected on chromosome arms 1BL, 2AL, 2BL, and 6BS, named as QYrXN3517-1BL, QYrXN3517-2AL, QYrXN3517-2BL, and QYrXN3517-6BS, respectively. Based on the Wheat 660 K array and bulked segregant exome sequencing (BSE-Seq), the most effective QTL on chromosome 1BL is most likely different from the known adult plant resistance gene Yr29 and was mapped to a 1.7 cM region [336 kb, including twelve candidate genes in International Wheat Genome Sequencing Consortium (IWGSC) RefSeq version 1.0]. The 6BS QTL was identified as Yr78, and the 2AL QTL was probably same as QYr.caas-2AL or QYrqin.nwafu-2AL. The novel QTL on 2BL was effective in seedling stage against the races used in phenotyping. In addition, allele-specifc quantitative PCR (AQP) marker nwafu.a5 was developed for QYrXN3517-1BL to assist marker-assisted breeding.