The team’s strategy rested on a classic tool of quantitative genetics: the recombinant inbred line population. Starting from a single cross between Svevo, an elite Italian durum wheat cultivar prized for its quality, and Zavitan, a wild emmer wheat accession originally collected in Israel, the researchers generated 135 recombinant inbred lines, or RILs. Each of these lines carries a unique mosaic of chromosome segments inherited from the two parents, produced through repeated self-pollination over multiple generations until the lines became genetically fixed. This means that every line is a stable, reproducible combination of Svevo and Zavitan DNA, allowing researchers to ask, for any given trait, which parental chromosome segments are associated with better performance.

To capture the traits that matter most for yield, the team measured four grain characteristics across four different environments: grain length, grain width, the ratio of length to width, and thousand-grain weight, a standard agronomic measure of how heavy an average sample of one thousand grains is. Evaluating the population in multiple environments is critical because grain traits are notoriously sensitive to growing conditions, and a genetic effect that appears in one field may vanish in another. Only loci that show consistent effects across environments, so-called stable quantitative trait loci or QTL, are truly useful for breeding, since breeders need markers that will deliver results regardless of where or when a variety is grown.

Genotyping the 135 lines was accomplished with the wheat 90K single nucleotide polymorphism array, a platform that interrogates tens of thousands of genetic markers spread across the wheat genome. Wheat is a challenging genome to work with: durum wheat carries two subgenomes, designated A and B, each contributing seven chromosome pairs, and the genome is enormous by any standard. The SNP array allowed the researchers to construct a linkage map and then scan the genome systematically for regions where the inheritance pattern of DNA markers correlated with the measured grain traits. Statistical mapping, using established QTL detection methods with empirical significance thresholds, then pinpointed the chromosomal intervals most likely to harbor genes influencing each trait.

The scan yielded eleven stable QTL, each detected consistently across the four test environments. These intervals ranged in physical size from 0.7 to 15.6 megabases on the durum wheat reference genome, a span that in some cases is small enough to narrow down candidate genes with reasonable confidence. Three of the stable loci, named QGLsv.swust-5AL, QGLsv.swust-5BL, and QGLsv.swust-7AL, influenced grain length; two, QGWsv.swust-4AL and QGWsv.swust-5AL, influenced grain width; and two, QTGWsv.swust-2BL and QTGWsv.swust-5BL, influenced thousand-grain weight. For all seven of these loci, the favorable allele, meaning the version of the gene region associated with larger or heavier grains, came from the cultivated parent Svevo, reflecting the cumulative effect of decades of selection by durum breeders.

The most intriguing finding, however, concerned the remaining four loci. All four QTL for grain length-width ratio, designated QLWRzv.swust-2BS, QLWRzv.swust-4AL, QLWRzv.swust-4BL, and QLWRzv.swust-6AS, carried their positive alleles from Zavitan, the wild parent. Grain shape, as captured by the length-to-width ratio, matters for both milling performance and market quality in durum wheat, and the fact that the wild ancestor consistently contributed the favorable alleles at these loci demonstrates that domestication did not exhaust the wild gene pool’s value. Wild emmer wheat, which grows naturally in the Fertile Crescent and has survived thousands of years of environmental fluctuation without human intervention, retains genetic variation that modern cultivars have lost, and this study shows that some of that variation is directly relevant to yield-related traits.

Quantitatively, the eleven stable QTL each explained between 9.22 and 19.22 percent of the phenotypic variation in their respective traits. These are substantial effects for complex agronomic traits, which are typically controlled by many genes of small effect. The fact that individual loci account for nearly a fifth of the variation in a trait like grain weight suggests that marker-assisted selection targeting these regions could produce measurable improvements in a breeding program. Marker-assisted selection, in which breeders use DNA markers rather than visible traits to track the inheritance of favorable genes, is particularly powerful for traits like grain weight that are difficult or expensive to measure directly on large numbers of early-generation plants.

To move from statistical associations to biological understanding, the researchers projected their QTL intervals onto three available reference genomes: Chinese Spring, the standard reference for bread wheat, and the Svevo and Zavitan genome assemblies for durum wheat. This cross-genome comparison identified 67 high-confidence candidate genes within the QTL intervals on the Chinese Spring reference, 88 on the Svevo reference, and 70 on the Zavitan reference. Functional annotation of these candidate genes revealed that they are enriched for roles in processes that make intuitive sense for grain development: seed development itself, transcriptional regulation, hormone signaling pathways, ubiquitin-mediated protein degradation, and carbohydrate metabolism. Each of these processes has well-documented connections to how a developing grain accumulates starch and protein and how large it ultimately grows.

The candidate gene list connects to a growing body of wheat functional genomics. Recent studies have identified genes such as TaSWEET11 and TaSWEET13h, sucrose transporters that are critical for grain filling, transcription factors like TaDOF6 that regulate sugar and gibberellin transport into the endosperm, and cytochrome P450 genes of the CYP78A family that influence seed size through hormone-mediated pathways. Several of the candidate genes identified in the new study fall into these same functional categories, suggesting that the mapped QTL may harbor orthologs or paralogs of genes whose effects have already been validated in bread wheat and rice. This convergence strengthens confidence that the mapped loci are not statistical artifacts but genuine biological regulators of grain development.

The practical implications extend beyond basic science. The stable QTL and their linked SNP markers constitute a direct resource for marker-assisted breeding in durum wheat, allowing breeders to introgress favorable alleles from wild emmer into elite backgrounds while simultaneously tracking the cultivated alleles that boost grain size and weight. Because the QTL were validated across four environments, breeders can have reasonable confidence that these markers will perform consistently across the diverse conditions in which durum wheat is grown, from the rain-fed fields of the Mediterranean to irrigated plains elsewhere. More broadly, the study reinforces a central lesson of modern crop improvement: the wild relatives of our staple crops are not relics of the past but living repositories of genetic variation, and interspecific crosses that tap into them can enrich the genetic basis of yield in ways that within-crop selection alone cannot achieve. As the global population grows and climate change pressures wheat production systems, tools like these that bridge the gap between wild genetic resources and elite cultivars will become increasingly indispensable.

Subject of Research: Mapping stable quantitative trait loci for grain yield-related traits in a durum wheat Svevo × Zavitan recombinant inbred line population

Article Title: Identification of stable QTL for yield-related traits in a durum wheat Svevo × Zavitan RIL population

Article References: Zhou, X., Zhou, B., Zhang, G., Zou, G., Yang, X., Xia, C., Li, X., Zheng, S., & Ren, Y. (2026). Identification of stable QTL for yield-related traits in a durum wheat Svevo × Zavitan RIL population. Theoretical and Applied Genetics, 139(10), Article 275. https://doi.org/10.1007/s00122-026-05389-1

Image Credits: AI Generated

DOI: 10.1007/s00122-026-05389-1

Keywords: durum wheat, wild emmer, QTL mapping, grain weight, grain size, recombinant inbred lines, marker-assisted breeding, SNP array, candidate genes, crop yield, genetic improvement, Triticum turgidum

At the heart of the work is a fundamental challenge in additive manufacturing: the properties of a 3D-printed metal part depend intimately on its grain structure, the spatial arrangement and orientation of crystalline regions separated by grain boundaries. Fine grains typically enhance strength and toughness through grain-boundary strengthening, while larger or anisotropic grains can improve ductility and creep resistance. Controlling these attributes during printing requires understanding how processing conditions such as laser scanning velocity, melt pool geometry, and heat-affected zone dimensions translate into grain evolution, a relationship complicated by vast multidimensional parameter spaces and non-linear thermal dynamics.

The research team, led by Dipayan Sanpui and Subramanian K.R.S. Sankaranarayanan, built their framework on a repository of 1,524 successfully retrieved three-dimensional polycrystalline microstructures generated with the open-source SPPARKS kinetic Monte Carlo code. These simulations modeled hundreds of laser passes over a substrate using a modified Potts Monte Carlo approach, capturing curvature-driven grain growth, melting, remelting, and nucleation in the heat-affected zone. Rather than relying on abstract statistical descriptors such as two-point statistics or chord length distributions used in earlier studies, the team extracted physically interpretable features directly from two-dimensional slices of the simulated structures.

Four descriptors formed the quantitative vocabulary of the study: average grain size, surface-to-volume ratio, sphericity, and roundness. Grain size serves as a first-order measure linked directly to mechanical performance through the Hall-Petch relationship, while surface-to-volume ratio distinguishes microstructures with similar grain dimensions but different boundary complexity. Sphericity captures overall grain compactness and elongation, whereas roundness reflects the sharpness of grain edges and corners, quantities that differ when two grains share similar elongation but diverge in boundary smoothness due to solidification dynamics. The team computed these measures directly from grain identifiers in the SPPARKS output files, streamlining integration into automated materials design pipelines.

With the features extracted, the researchers framed the problem as supervised regression, using seven processing conditions, including scanning velocity, scanning pattern, molten zone width, depth, and tail length, and heat-affected zone width and tail length, as inputs. They trained and compared multiple algorithms: Random Forest, Gradient Boosting Machine, Extreme Gradient Boosting (XGBoost), a feed-forward neural network, and a one-dimensional convolutional neural network, with hyperparameters optimized through six-fold grid search cross-validation and Keras-Tuner’s Hyperband search for the neural architectures.

Tree-based ensemble methods consistently outperformed the neural networks. XGBoost emerged as the best overall performer, achieving a coefficient of determination of 0.977 for grain size prediction, 0.946 for surface-to-volume ratio, 0.954 for sphericity, and 0.957 for roundness on test data, with a mean absolute error of just 6.62 pixels for grain size. The authors attribute grain size’s high predictability to its status as a first-order reflection of the temperature distribution across the simulation domain, while shape descriptors such as sphericity and roundness depend on secondary morphological effects, including grain boundary curvature and irregular grain interactions, that are harder to infer from processing conditions alone.

A particularly striking finding emerged from the parity plots: grain size predictions split into two distinct clusters corresponding to different microstructural regimes. The first cluster, with average grain sizes between 0 and 450 pixels, represented anisotropic structures of alternating equiaxed and elongated grains formed at higher scanning velocities and broader heat-affected zones. The second cluster, spanning 600 to 1200 pixels, arose from low-velocity, narrow-heat-affected-zone conditions that promote localized remelting and coarser, irregular grains. When trained and evaluated within the well-behaved first regime, the model reached an R² of 0.982 with an error of only 2.305 pixels; in the morphologically heterogeneous second regime, accuracy dropped to R² of 0.561, revealing where the surrogate’s reliability limits lie.

To open the black box, the team applied Shapley Additive exPlanations (SHAP), a game-theory-based framework that quantifies each input feature’s marginal contribution to predictions. The analysis showed that scanning velocity and heat-affected zone characteristics dominate grain size outcomes globally, consistent with established solidification theory in which thermal gradients, cooling rates, and grain-boundary mobility govern grain evolution. At high velocities, scanning speed and HAZ tail length controlled the outcome, with rapid cooling producing fine, anisotropic grains. At minimum velocity, scanning direction and HAZ tail length became the primary drivers: cross-hatch patterns induced multidirectional heat flux and rapid cooling that favored clusters of fine equiaxed grains, while unidirectional scanning promoted elongated, directionally aligned grains.

Beyond interpretation, the framework demonstrated practical utility for process optimization. By filtering candidate parameter combinations according to the relative error between predicted and desired grain sizes, the researchers identified viable alternative processing conditions under fixed maximum or minimum laser velocities. Under high velocity, neighboring points in parameter space with variations in melt-zone geometry yielded nearly identical grain structures, suggesting flexibility in achieving consistent microstructures. Under low velocity, HAZ tail length and scanning direction offered the most meaningful degrees of freedom. The computational payoff is substantial: simulating a single 100 × 100 × 100 site microstructure took roughly 25 minutes across four compute nodes, while the trained models deliver predictions in minutes once trained.

The authors envision their approach as a foundation for digital twins and closed-loop process control, in which sensor-derived process variables feed continuously into trained models for online quality monitoring and adaptive parameter optimization. They acknowledge limitations, notably that featurization currently operates on two-dimensional slices rather than full three-dimensional volumes, and they point toward extending the framework to experimental microscopy images, inverse design, and active-learning-based calibration as next steps. By coupling high-fidelity physics simulation with interpretable machine learning, the study illustrates how explainable AI can transform additive manufacturing from a process of empirical trial and error into a predictive, quantitatively grounded engineering discipline.

Subject of Research: Interpretable machine learning prediction of process–structure relationships in metal additive manufacturing using kinetic Monte Carlo microstructure simulations

Article Title: Data-driven discovery of process–structure relationships in additive manufacturing via featurization from kinetic Monte Carlo simulations and interpretable machine learning

Article References: Sanpui, D., Chan, H., Koneru, A., Banik, S., Manna, S., & Sankaranarayanan, S. K. (2026). Data-driven discovery of process–structure relationships in additive manufacturing via featurization from kinetic Monte Carlo simulations and interpretable machine learning. Machine Learning with Applications, 25, Article 100985. https://doi.org/10.1016/j.mlwa.2026.100985

Image Credits: AI Generated

DOI: 10.1016/j.mlwa.2026.100985

Keywords: additive manufacturing, kinetic Monte Carlo, machine learning, microstructure, XGBoost, SHAP, grain size, process-structure linkages, featurization, surrogate modeling, explainable AI, metal 3D printing