Wild Relatives Hold the Genetic Keys to Future-Proof Crops, Review Finds

Crop wild relatives, the untamed ancestors and cousins of modern food crops, have long been touted as a treasure trove of genetic diversity for plant breeding. Yet a sweeping new review in Theoretical and Applied Genetics argues that the field has been seduced by technological promise at the expense of genetic realism. Ruixue Xiao and Xiaopeng Li, researchers at the Zhengzhou Fruit Research Institute of the Chinese Academy of Agricultural Sciences, contend that translating wild diversity into better cultivars is blocked not primarily by a lack of tools, but by fundamental quantitative-genetic and population-genetic bottlenecks that the breeding community has yet to confront head-on.

The core of their argument rests on a sobering historical fact: domestication was a massive genetic filter. As humans selected a handful of preferred plants from wild populations, allele after allele was lost, a phenomenon vividly documented in wheat, where genome-wide surveys have shown a dramatic erosion of nucleotide diversity since domestication. This so-called domestication bottleneck did not merely delete variation; it reshaped the architecture of the genome itself. Linkage disequilibrium, the non-random association of alleles along chromosomes, became extended in elite breeding pools, meaning that breeders selecting for one trait inevitably drag along large blocks of linked DNA, both desirable and deleterious. The result is that modern crops carry a fraction of the allelic richness of their wild progenitors, and the variation that remains is organized in ways that complicate further improvement.

Wild relatives, by contrast, retain extensive diversity accumulated over millions of years of evolution in diverse environments, including alleles conferring resistance to pests and pathogens, tolerance to drought, heat, salinity, and flooding, and novel nutritional profiles. The review synthesizes case studies across cereals, legumes, and Solanaceous crops to illustrate both the promise and the pitfalls. Classic successes include the Sub1 locus from wild rice, which confers submergence tolerance and has been deployed in flood-prone rice paddies across Asia, and resistance genes such as Lr67/Yr46 in wheat, which provides broad-spectrum, adult-plant resistance to rust and powdery mildew. In tomato, decades of work with introgression line populations derived from the wild relative Solanum pennellii have mapped yield-associated quantitative trait loci and even revealed wild alleles that boost fruit soluble solids, a trait once thought to be exhausted in cultivated germplasm.

But these triumphs mask a deeper problem that the authors place at the center of their framework: not every wild allele transfers cleanly into a cultivated background. They distinguish between loci with stable, reproducible breeding effects and those confined to proof-of-concept discovery. Quantitative genetics offers a mechanistic explanation for this divide. The effect of an allele depends on its genetic background, a phenomenon known as epistasis, in which genes interact with one another rather than acting independently. An allele that boosts yield in a wild genome may do nothing, or even harm, when moved into an elite cultivar carrying a different constellation of interacting genes. Linkage drag compounds the problem: when breeders cross a wild relative with a crop, they transfer not just the target gene but thousands of neighboring wild genes, many of which carry deleterious effects accumulated in the wild. Repeated backcrossing can shrink these blocks, but the process is slow, expensive, and often leaves unwanted wild DNA lurking near the target locus.

The review also scrutinizes the population-genetic limitations of the emerging biotechnological arsenal. Pan-genomics, which sequences hundreds or thousands of individuals to capture the full complement of genes within a species, has exploded in popularity. Graph-based pan-genomes of rice, soybean, tomato, chickpea, and Brassica crops have revealed structural variation, presence-absence variants, and massive haplotypes that single reference genomes missed entirely. Super-pangenomes that integrate wild species alongside cultivated ones promise an even more complete catalog of usable variation. Yet the authors caution that cataloging variation is not the same as deploying it. A pan-genome can identify a wild haplotype associated with heat tolerance, but predicting how that haplotype will behave in an elite background, and whether its effect will be large enough to matter in a farmer’s field, requires quantitative-genetic models that remain underdeveloped.

Genome editing faces parallel constraints. CRISPR-Cas9 has enabled striking demonstrations of de novo domestication, in which researchers edit wild species to recapitulate key domestication traits. Wild tomato relatives have been edited to produce compact, productive plants with larger fruits, and orphan crops have been rapidly improved by targeting genes controlling shoot architecture and flowering. Editing cis-regulatory elements has even allowed fine-tuning of quantitative trait variation, offering a way to sculpt trait expression rather than simply knocking genes out. However, the review emphasizes that these successes have largely targeted major-effect genes with well-understood functions. Most agronomically valuable traits, including yield, stress tolerance, and nutrient use efficiency, are highly polygenic, controlled by hundreds of loci of small effect scattered across the genome. Editing such traits one gene at a time is impractical, and delivery of editing machinery into many wild species remains inefficient because transformation protocols are species-specific and often poorly optimized outside major crops.

Genomic selection, which uses genome-wide markers to predict the breeding value of individuals, offers another route to harnessing wild diversity. The method has delivered measurable genetic gains in maize breeding programs, including under drought stress, and has been proposed as a way to turbocharge gene banks by predicting which accessions harbor useful alleles before they are ever field-tested. But prediction accuracy depends on the relatedness between the training population and the selection candidates, and on the extent of linkage disequilibrium between markers and causal variants. When the training population is elite germplasm and the candidates are wild or landrace material, the relationships are so distant that predictions can collapse. The authors argue that pre-breeding programs, which cross wild material with elite lines to create bridging populations such as MAGIC and nested association mapping populations, are essential to make wild alleles predictable and selectable.

What sets this review apart is its insistence on testable hypotheses rather than technological optimism. The authors propose a mechanistic roadmap for prioritizing CWR resources: first, characterize the architecture of the domestication bottleneck in each crop to understand what variation was lost and how it is organized; second, distinguish wild loci whose effects are stable across backgrounds from those that are background-dependent, using designed populations and multi-environment testing; and third, evaluate each biotechnology against the population-genetic realities of the target species before investing in deployment. This framework shifts the discourse from what genome editing and pan-genomics could theoretically do to what the underlying genetics will actually permit, a reframing the authors argue is overdue in a field awash with hype.

The stakes could hardly be higher. Global agriculture faces converging pressures from climate change, which is intensifying droughts, heat waves, and disease outbreaks faster than conventional breeding can respond, and from ongoing genetic erosion, which is depleting both crop diversity and the wild ecosystems that harbor it. Many crop wild relatives are themselves threatened in their native habitats, making their collection, conservation, and characterization an urgent priority. The review’s message is ultimately one of disciplined hope: the wild gene pool remains the single largest reservoir of untapped adaptive variation available to breeders, but unlocking it will require confronting epistasis, linkage drag, and prediction limits with the same rigor that has been devoted to developing new tools. For a world that must feed ten billion people on a warming planet, the wild cousins of our crops may be the most valuable untapped resource in agriculture, provided breeders and geneticists learn to read their genomes with quantitative-genetic realism rather than technological faith.

Subject of Research: Use of crop wild relatives and their genetic diversity in modern crop improvement and breeding

Article Title: Unlocking the future of crop improvement with crop wild relatives

Article References: Unlocking the future of crop improvement with crop wild relatives. (n.d.). https://doi.org/10.1007/s00122-026-05399-z

Image Credits: AI Generated

DOI: 10.1007/s00122-026-05399-z

Keywords: crop wild relatives, plant breeding, domestication bottleneck, pan-genomics, genome editing, genomic selection, quantitative genetics, linkage disequilibrium, epistasis, de novo domestication, genetic diversity, food security