SNP Array Reveals Hidden Genetic Wealth in Repatriated Indian Wheat Landraces

A vast trove of old wheat seeds, gathered from farmers’ fields across India decades ago and only recently brought back home, is turning out to hold far more genetic diversity than many breeders had assumed. In a study published in BMC Plant Biology, a team at the ICAR-National Bureau of Plant Genetic Resources in New Delhi, working with colleagues at the Borlaug Institute for South Asia and the ICAR-Indian Agricultural Research Institute, used a high-throughput SNP genotyping array to chart the molecular diversity of 828 wheat landraces alongside 100 modern cultivars. Their analysis not only maps the population structure of this repatriated germplasm in unprecedented detail but also distills it into a compact core collection that preserves nearly all of the genetic variation of the full set while requiring breeders to manage only a quarter of the accessions.

The stakes of this work are considerable. Wheat is a staple crop for billions of people, and the pressures of climate change, heat stress, water scarcity, and a growing global population are demanding new sources of genetic variation that modern breeding programs can deploy. Landraces, the traditional farmer-selected varieties that predate formal breeding, are widely recognized as reservoirs of adaptive alleles lost during the intensification of agriculture. Yet gene bank collections are often so large and so poorly characterized at the molecular level that breeders struggle to use them efficiently. Repatriation efforts, in which material collected from a country decades ago is returned from international gene banks, have restored substantial holdings to national programs, but turning those holdings into usable breeding resources requires a clear picture of what diversity they actually contain.

The research team genotyped the 928 accessions using array-based single nucleotide polymorphism markers, a technology that interrogates hundreds of thousands of known variant sites across the genome in a single assay. Because bread wheat is hexaploid, carrying three related subgenomes descended from different ancestral grass species, the researchers could examine diversity patterns separately in the A, B, and D genomes. The B genome emerged as the most densely covered with SNPs, followed by the A and then the D genome. At the chromosome level, 2B carried the highest marker density while 4D carried the lowest, a pattern that mirrors what is already known about the evolutionary history of hexaploid wheat, in which the D genome contributed by the wild goatgrass Aegilops tauschii retains comparatively less sequence variation because it entered the wheat lineage most recently.

Genome-wide diversity statistics told a nuanced story. Gene diversity, also known as expected heterozygosity, and the polymorphism information content, a measure of how informative a marker is for detecting genetic differences, were both slightly higher in the B and D genomes than in the A genome. The team also computed Tajima’s D, a statistic that compares the number of segregating sites with the average pairwise differences between sequences and can signal whether a population has undergone recent selection or demographic change. The slightly negative values observed, interpreted in a comparative framework, were consistent with a relatively stable demographic history and weaker recent selection pressure in the landraces compared with the improved cultivars. Notably, nucleotide diversity was lower in the landraces than in the improved genotypes, a finding the authors interpret carefully rather than as a simple claim that landraces are uniformly more variable.

A key quality-control step involved examining the ratio of transitions, mutations between the two purines or the two pyrimidines, to transversions, mutations that swap a purine for a pyrimidine. In biological data, transitions typically occur more frequently than transversions, and genotyping arrays are designed with this bias in mind. The Ts/Tv ratios exceeded two across all three subgenomes, supporting the reliability of the SNP set and indicating that the marker data were not dominated by technical artifacts. This kind of validation matters enormously in diversity studies, because spurious markers can distort downstream analyses of population structure and relationships, leading researchers astray about the true genetic architecture of a collection.

To uncover how the accessions are organized genetically, the researchers deployed two complementary clustering approaches. STRUCTURE, a Bayesian clustering algorithm that assigns individuals to populations based on allele frequencies, identified two major genetic clusters when the number of clusters, K, was set to two. In parallel, discriminant analysis of principal components, or DAPC, using the Bayesian Information Criterion to choose the optimal number of groups, inferred six clusters. The authors reconcile these results by describing a hierarchical population structure: the primary level of genetic differentiation in the collection splits the germplasm into two broad groups, while progressively higher values of K reveal finer-scale subdivision within those groups. This layered organization is typical of germplasm assembled from many regions and eras of cultivation, and it gives breeders a practical map for sampling diversity at whatever resolution their objectives require.

An analysis of molecular variance, or AMOVA, added an important caveat to the clustering picture. Most of the genetic variation in the collection resides within groups rather than between them, meaning that any single landrace population already contains a substantial share of the total diversity. For breeders, this implies that carefully chosen individuals from within a cluster can capture much of what the cluster has to offer, and it underscores why a well-constructed core set, rather than a simple geographic sample, is the most efficient route into the collection. The team also examined linkage disequilibrium, the non-random association of alleles at different loci, and found that LD decayed fastest in the A genome and slowest in the B genome. LD decay rates shape the resolution of genome-wide association studies: faster decay means a marker must sit closer to the causal gene to show an association, so knowing the per-genome LD landscape helps researchers design mapping experiments in this material.

The centerpiece of the practical output is the core collection. Using a maximization strategy, an algorithmic approach that selects accessions to retain as many alleles as possible while minimizing redundancy, the team distilled the full panel down to 25 percent of the accessions. Remarkably, this core retained 99.74 percent of the alleles detected in the complete set and showed diversity statistics that were similar to, or in some measures greater than, those of the full collection. In effect, the core set functions as a compressed but nearly complete snapshot of the genetic variation in the repatriated landraces, allowing breeders and geneticists to screen for traits such as heat tolerance, disease resistance, or grain quality without genotyping and phenotyping hundreds of accessions first.

The study’s authors frame the results as evidence that repatriated Indian landraces contain significant and structured genetic variation, and they position the core set as a well-defined resource for future association studies, pre-breeding, and the development of climate-resilient wheat. Pre-breeding, the process of transferring useful alleles from unadapted germplasm into breeding-ready material, is often bottlenecked by the sheer size and poor characterization of gene bank collections, and a validated core set directly addresses that bottleneck. The work was funded by the Department of Biotechnology, Government of India, under a project on germplasm characterization and trait discovery in wheat using genomics approaches, with additional support from the Indian Council of Agricultural Research, reflecting a national commitment to converting archived diversity into modern agricultural value.

Beyond its immediate utility for wheat improvement, the study offers a template for how national gene banks worldwide can handle repatriated material. The combination of array-based genotyping, hierarchical structure analysis, and maximization-based core construction provides a reproducible pipeline for transforming a static collection into an active research platform. As sequencing costs fall and reference genomes improve, the SNP framework established here can be extended with haplotype-level analyses and genome-wide association scans to pinpoint the alleles underlying traits that landrace farmers selected for over centuries, including tolerance to marginal soils, variable rainfall, and endemic pathogens. In an era when crop genetic erosion continues and climate volatility accelerates, the message of this research is that the seeds sitting in cold storage are not relics but a living, structured archive, and with the right molecular tools, that archive can be opened efficiently enough to matter for the wheat fields of the coming decades.

Subject of Research: Genetic diversity and population structure of repatriated Indian wheat landraces analyzed with SNP array markers to develop a core collection

Article Title: Molecular diversity analysis and development of a core set in wheat using array-based SNP markers

Article References: Singh, V., Shree, B., Bhaskar, Y., Singh, A., Kumari, J., Thakur, S., Bhati, P. K., Singh, C. K., Jacob, S. R., Singh, S. K., Yadav, M. C., Singh, R., Kumar, S., & Singh, A. K. (2026). Molecular diversity analysis and development of a core set in wheat using array-based SNP markers. BMC Plant Biology. https://doi.org/10.1186/s12870-026-09969-7

Image Credits: AI Generated

DOI: 10.1186/s12870-026-09969-7

Keywords: wheat, landraces, SNP markers, genetic diversity, core collection, population structure, Triticum aestivum, germplasm, linkage disequilibrium, Tajima's D, pre-breeding, climate resilience