The research, led by Le Wang and Yuan Zhang of the Northeast Institute of Geography and Agroecology at the Chinese Academy of Sciences, together with corresponding author Hengyou Zhang, began with a panel of 450 soybean accessions collected from northern China. This kind of diversity panel is the raw material of a genome-wide association study, or GWAS, an approach that scans the genomes of many individuals for genetic variants that correlate with a measurable trait. In this case, the trait was leaf shape, a characteristic that varies visibly across soybean germplasm from slender, narrow leaflets to broad, rounded ones. The team scored the plants across two consecutive growing seasons, a repetition that helps distinguish genuine genetic signals from the noise of weather, soil, and measurement error.

The statistical machinery behind the analysis was deliberately conservative. Rather than relying on a single association model, the researchers ran multiple complementary GWAS models, including mixed-model frameworks that account for the relatedness among accessions and the population structure that inevitably arises when plants are drawn from different geographic regions. A signal that survives this gauntlet—appearing consistently across models and across years—carries far more weight than one that flickers in and out of significance. The robust quantitative trait locus that emerged, which the team designated qLS, sits on chromosome 20 and was detected reliably under every analytical condition the researchers applied.

Within that locus, the strongest association pointed to a single nucleotide polymorphism, labeled SNP_35828042, located at position 35,828,042 on the chromosome. What makes this variant especially compelling is where it lands: inside the second exon of a gene called GmJAG1, which encodes a C2H2-type zinc finger transcription factor. The gene is a soybean homolog of JAGGED, a well-studied regulator of lateral organ growth in Arabidopsis thaliana, the mustard weed that has long served as genetics’ favorite reference plant. Because the SNP falls in an exon, it is not a silent marker sitting near the causal gene—it is a change in the gene’s own protein-coding sequence.

That change is a G-to-C substitution, and it is not benign. The single-letter swap converts a codon so that aspartic acid is replaced by histidine at the ninth amino acid position of the protein, an alteration the authors denote D9H. Critically, this substitution occurs within the EAR motif, a short amino acid sequence that is highly conserved across plants and serves as a transcriptional repression domain. EAR-motif-containing repressors are central players in plant development, acting as molecular switches that silence target genes and thereby shape organs as they grow. Altering even one residue in such a conserved motif can change how strongly the protein represses its targets, and the narrow-leaf-associated C allele appears to do exactly that, shifting the developmental program toward narrower leaflets.

Association studies can only suggest causation, so the team turned to independent functional evidence. Previous work had generated soybean plants in which GmJAG1 was knocked out using CRISPR-Cas9 gene editing, and analysis of those edited lines supported the gene’s role in regulating leaf shape. The new population-scale data thus dovetail with direct experimental manipulation: natural variation at GmJAG1 tracks with leaf shape across hundreds of accessions, and disrupting the gene changes leaf morphology in edited plants. Together, these lines of evidence elevate GmJAG1 from a statistical hit to a functionally supported target for breeding.

One of the most striking findings is geographic. When the researchers examined how the two alleles are distributed across the collection, they found that the narrow-leaf-associated C allele becomes more frequent in accessions from more northern latitudes. This northward enrichment is consistent with regional differentiation, suggesting that the variant has been shaped by selection or breeding history as soybean adapted to shorter growing seasons and different light environments. Narrow leaves may confer advantages in high-latitude conditions, where dense planting and rapid canopy closure can maximize the capture of limited seasonal sunlight. The pattern offers a rare glimpse of how a single coding change in a developmental regulator mirrors the geographic mosaic of crop adaptation.

The practical payoff of the study is a diagnostic marker. The researchers developed a derived cleaved amplified polymorphic sequence marker—a dCAPS marker—targeting the G/C polymorphism. dCAPS is an elegant workaround for variants that do not naturally create or destroy a restriction enzyme cutting site: a mismatched base is introduced into one of the PCR primers, so that amplification creates an artificial restriction site for one allele but not the other. Digesting the PCR product with the appropriate enzyme then yields fragment patterns that reveal which allele a plant carries. When the team tested the marker on 41 accessions, the genotypes showed complete concordance with visual leaf-shape classification—every plant called narrow by the marker looked narrow, and every plant called broad looked broad.

That level of agreement matters because it transforms leaf shape from a trait that must be judged by eye, often late in the season and subject to environmental influence, into a DNA test that can be run on a seedling in days. Marker-assisted selection built on such diagnostic markers allows breeders to screen thousands of seedlings cheaply, discard the ones lacking the desired allele, and devote field space and time only to plants carrying the genetic architecture they want. In soybean, where canopy architecture influences light interception, photosynthesis, and ultimately yield, a reliable marker for leaf shape slots directly into programs pursuing the kind of plant-ideotype optimization that drove yield gains in rice and wheat during past Green Revolutions.

The study also fills a gap in soybean genetics. A major leaflet-shape gene, Ln, was cloned years ago and is known to pleiotropically influence both leaf shape and seed number per pod, but the natural variation underlying leaf shape across diverse germplasm has remained underexploited for molecular breeding. By validating GmJAG1 as a leaf-shape locus at population scale, characterizing its geographic distribution, and delivering a working dCAPS assay, the team has added a second, independently actionable handle on the trait. For a crop that supplies the majority of the world’s protein meal and a large share of its vegetable oil, and whose yields must rise on finite land under a changing climate, turning a conserved zinc finger transcription factor into a breeder’s test tube assay is the kind of quiet, cumulative advance on which agricultural progress actually depends.

Subject of Research: Genetic basis of leaf shape variation in soybean and development of a diagnostic molecular marker for breeding

Article Title: High-resolution GWAS and diagnostic functional marker development for leaf shape in soybean

Article References: High-resolution GWAS and diagnostic functional marker development for leaf shape in soybean. (n.d.). https://doi.org/10.1007/s00122-026-05408-1

Image Credits: AI Generated

DOI: 10.1007/s00122-026-05408-1

Keywords: soybean, leaf shape, GWAS, GmJAG1, dCAPS marker, marker-assisted selection, quantitative trait locus, canopy architecture, zinc finger transcription factor, EAR motif, crop breeding, chromosome 20

The team set out to dissect the genetic architecture of internode number and internode length, the two traits that together define plant height in rice. Rather than relying on classical crosses between two parents, the researchers deployed a genome-wide association study, or GWAS, across a remarkably diverse panel of 429 rice accessions. This population-scale approach exploits the natural historical recombination of the rice genome: if a particular DNA variant consistently tracks with shorter internodes across hundreds of unrelated varieties, that variant likely sits near a gene genuinely involved in the trait. The scale of the genotyping was impressive in its own right—3.16 million high-quality single nucleotide polymorphisms were used to scan the twelve rice chromosomes for statistical associations with each measured trait.

The scan paid off. The analysis identified 34 significant marker-trait associations linking specific genomic regions to internode characteristics, and among these the researchers pinpointed four quantitative trait loci, or QTLs, that had not been described before. Within the most compelling regions, two candidate genes emerged and were christened Internode number and length 1 (INL1) and INL2. Functional annotations suggested that INL1 encodes a cytochrome P450 enzyme, while INL2 encodes a zinc finger protein transcription factor. Both gene families are heavyweight players in plant biology: cytochrome P450s catalyze an enormous range of biochemical reactions, including steps in hormone synthesis, and zinc finger transcription factors act as molecular switches that regulate suites of downstream genes.

The choice of these two gene families is particularly intriguing given what is already known about stem elongation. Gibberellins, a class of phytohormones, are the principal drivers of internode elongation, and the Green Revolution sd1 gene itself encodes a gibberellin biosynthesis enzyme. Cytochrome P450 enzymes have repeatedly surfaced in this arena—EUI1, another rice P450 monooxygenase, modulates gibberellin responses to control internode elongation, and several P450s participate in brassinosteroid biosynthesis, a hormone pathway that crosstalks extensively with gibberellins. Similarly, Cys2/His2 zinc finger proteins in rice have been shown to repress SD1/OsGA20ox2, tying transcriptional control directly to gibberellin production. INL1 and INL2 therefore fit neatly into an emerging picture of height regulation as a multi-layered network combining hormone metabolism and transcriptional governance.

Correlation, however, is not causation, and the team moved to functional validation using CRISPR/Cas9 genome editing. By knocking out INL1 and INL2 in rice, the researchers created plants in which these genes no longer function. The result was unambiguous: the knockout lines displayed shorter internodes and reduced overall plant height compared with wild-type plants. This loss-of-function phenotype provides strong experimental evidence that both genes act as positive regulators of internode elongation—when they are disabled, stems grow shorter. For breeders, genes that reduce height when knocked out are exactly the kind of raw material needed for dwarf breeding programs aimed at improving lodging resistance.

To understand how the knockouts produce their dwarfed stature, the researchers turned to transcriptomics, measuring genome-wide gene expression changes in the edited plants. The analysis revealed that loss of INL1 and INL2 differentially modulates gibberellin metabolism, ultimately leading to the dwarf phenotype. In other words, disrupting these genes perturbs the hormonal machinery that normally drives cell elongation in the stem, choking back the growth of internodes. This places the two new genes upstream of, or within, the gibberellin pathway and suggests that they coordinate the metabolic flux that fuels stem extension.

Beyond gene function, the study explored natural allelic variation in INL1 and INL2 through haplotype analysis. Haplotypes are characteristic combinations of DNA variants inherited together at a locus, and different haplotypes often produce different trait values. The researchers found that accessions carrying the HapA type of the two genes had shorter internodes than those carrying other haplotypes. This means that favorable, height-reducing alleles of OsINL1 and OsINL2 already exist within the rice gene pool, circulating among natural and cultivated varieties. Rather than waiting for new mutations, breeders can in principle track these haplotypes with DNA markers and introgress them into elite cultivars through marker-assisted selection.

The practical implications extend well beyond plant architecture trivia. Lodging is a chronic threat to rice production worldwide, particularly as extreme weather events intensify and high-yielding varieties carry heavier panicles atop their stems. Lodged crops suffer reduced photosynthesis, impaired grain filling, greater disease pressure and harvest losses. Because plant height is determined by the number and length of individual internodes, genes that fine-tune each segment offer breeders granular control: reducing the length of specific internodes can lower the center of gravity and strengthen the culm without necessarily sacrificing the biomass or yield potential that taller plants sometimes confer. The authors note that the favorable alleles of INL1 and INL2 reduce height and can be further used in rice dwarf breeding to improve lodging resistance.

There is also a broader scientific payoff. Most dwarfing work in rice has centered on sd1 and a handful of related loci, and researchers have increasingly argued that additional dwarfing or semi-dwarfing genes are needed to diversify the genetic basis of height control and avoid the vulnerabilities of over-reliance on a single locus. By combining a large, diverse association panel, dense SNP coverage, new QTL discovery, CRISPR validation and transcriptomic profiling, this study demonstrates a full pipeline from statistical signal to mechanistic hypothesis to breeding-ready allele. The finding that a cytochrome P450 and a zinc finger transcription factor converge on gibberellin metabolism enriches the known regulatory map of stem growth and may illuminate how gibberellin homeostasis is tuned in other cereal crops such as wheat and barley, where similar architecture principles apply.

As global demand for rice continues to climb against a backdrop of shrinking arable land and volatile weather, the genetics of stem architecture has become a frontline topic in crop science. The identification of INL1 and INL2 shows how modern genomics—millions of SNPs, hundreds of accessions, precision gene editing and hormone-level transcriptomics—can convert a field-measurable trait like internode length into molecular tools that breeders can deploy. If the favorable haplotypes of these two genes perform as hoped when pyramided into elite backgrounds, the humble rice stalk may soon stand a little shorter, and a great deal steadier, in the fields that feed the world.

Subject of Research: Genetic basis of internode length and plant height regulation in rice

Article Title: Candidate genes identification for internode length in rice via genome-wide association studies

Article References: Li, B., Sun, J., Yao, Q., Ding, Y., Geng, Y., Bao, Y., Ang, Y., Jiang, H., Yang, Z., Shi, Y., Jiang, J., Dang, X., Zhang, C., Ergashev, M. A., Li, M., Ali, J., Li, Z., & Liu, E. (2026). Candidate genes identification for internode length in rice via genome-wide association studies. Theoretical and Applied Genetics, 139(10), Article 294. https://doi.org/10.1007/s00122-026-05395-3

Image Credits: AI Generated

DOI: 10.1007/s00122-026-05395-3

Keywords: rice, genome-wide association study, internode length, plant height, INL1, INL2, gibberellin, CRISPR/Cas9, lodging resistance, dwarf breeding, cytochrome P450, zinc finger transcription factor