AI could help breeders find climate-ready traits in wild crop relatives

A review describes how combining genetic maps and deep-learning systems may uncover useful variants missed by standard crop prediction.

Editorial illustration — not from the study.

Crop domestication narrowed the genetic diversity used in modern varieties. The review explains that important differences may remain in wild relatives, including structural variations such as DNA segments that are present in some plants but absent in others, or occur in different copy numbers.

The authors describe a shift from a single linear genome reference to richer representations, including genome graphs and short sequence patterns. They discuss how graph-based neural networks and Transformer-based models could incorporate these variations, account for interactions between genetic changes, predict the effects of previously uncharacterized wild variants and assess how genetic effects vary across environments.

How AI reads wild variation

The article reviews computational approaches for recovering structural variations that standard models based mainly on single-letter DNA changes and linear reference genomes can miss. It contrasts genome graphs, which represent alternative evolutionary paths, with alignment-independent sequence features such as K-mers, or short DNA sequences. The authors synthesize how graph neural networks and Transformer-based models could use these data to capture complex genetic interactions, support prediction of uncharacterized wild alleles and improve the study of genotype-by-environment interactions. The abstract presents these approaches as a route toward using wild germplasm in breeding and accelerating de novo domestication of climate-resilient crops.

Why wild diversity matters

Wild crop relatives retain adaptive diversity that was reduced as crops were domesticated. Finding and predicting the effects of variants hidden in those relatives could broaden the genetic material available to breeders, particularly when traits depend on complex DNA changes or on the growing environment. The review suggests that combining AI with pangenomics—a way of representing genetic diversity across many genomes—may make that diversity more usable, but it does not establish that a particular crop variety or climate trait has already been improved.

Evidence and caveats

This is a review and synthesis of emerging computational methods, rather than a report of a single breeding trial or a dataset with measured crop outcomes. The abstract gives no sample sizes, benchmark results or specific crop examples, and it does not quantify the accuracy of the proposed predictions. Its discussion of zero-shot prediction and climate-resilient breeding describes potential applications; testing how reliably these methods predict real traits across wild and cultivated plants remains necessary.

Pulsating stars in binary systems can reveal what happens inside them

A review examines how pulsating stars in eclipsing binary systems can be used to study stellar interiors and the effects of tides. Space-based observations, combined with repeated measurements of starlight spectra, can provide model-independent masses and radii for thousands of systems.

12 hours ago·6 views

Physics & Space

Sound waves traveled steadily despite being confined to irregular spots

Researchers observed sound waves moving ballistically through a disordered acoustic lattice even though the system’s wave states were strongly localized at irregular positions. The result separates localization of wave states from the movement of wave packets.

19 hours ago·3 views

Physics & Space

A new way to map cell detail inside thick samples needs no added labels

Researchers developed an imaging approach that reconstructs label-free, three-dimensional subcellular detail in thick, highly scattering biological samples. Tests on model samples and biological specimens showed that specific measurement and calculation choices produced high-quality maps of the samples’ internal optical properties.

1 day ago·39 views