URBANA, Ill. — Small grains researcher Juan David Arbelaez-Velez knows the secret to making perfect rice — and it’s not about how you cook it. Arbelaez and his team are investigating the genetic blueprint that determines different grain attributes such as appearance, cooking time, and texture. Their paper, published in The Plant Genome, offers a strategy that will help breeders improve grain quality holistically, while cutting costs and saving time.
“Sometimes a variety is so productive that breeders will overlook some of its flaws. Then when you get to quality testing, the variety doesn’t meet the industry and consumer standards,” said Arbelaez, an assistant professor in the Department of Crop Sciences, part of the College of Agricultural, Consumer and Environmental Sciences at the University of Illinois Urbana-Champaign. “When we use multi-trait genetic approaches, we take advantage of all the available information — whole genome DNA, agronomic traits like yield, maturity, and plant height, in addition to secondary quality traits, like starch content, grain size, color, and appearance.”
Developing better rice
The global demand for high-quality rice is increasing, Arbelaez says, and rice breeders have to keep up. Growers want productive varieties that grow well and can survive drought and pests. Millers want rice grains that will survive the hulling and polishing processes without breaking, while consumers look for rice that will cook easily and have a pleasant texture.
The traditional crop breeding process can be slow, as crop productivity and grain quality are often assessed separately. Paying quality sensory evaluators can also get expensive. By predicting grain qualities beforehand using genomic selection, breeders can focus their efforts and resources on varieties that will perform well on the market.
Modern-day plant breeders often work on one gene at a time. For example, a particular gene might explain 40% of the variation in grain size between different rice varieties, so researchers will develop varieties with the version of that gene that yields desirable results. The problem, though, is the other 60%. Genes scattered throughout the entire genome, along with environmental conditions, can sometimes “overpower” the singular gene that codes for most of a trait, especially when minor changes determine consumer acceptance.
Harnessing the genome
That’s why the researchers combined a bird’s-eye view of the genome with a microscopic look at individual genes, using a method known as marker-assisted multi-trait genomic selection.
“The longer, less cost-effective route would be to study the impacts of a single allele,” Arbelaez said. “Instead, we’re looking at the whole genome, and several different variations.”
The basis of their technique, multi-trait genomic selection (MT-GS), investigates the big picture. Instead of looking at individual genes that code for individual traits — what geneticists have usually done — MT-GS predicts multiple traits at once based on the whole genome, or the entirety of an organism’s genetic code.
To make their MT-GS model more accurate, the researchers zoomed in on individual genetic markers with known links to certain traits, such as texture and cooking time. This is the “marker-assisted” part of their technique. It clarifies the picture for the model, allowing it to “learn” from known genetic relationships and identify linked traits.
“In genomic selection, we're looking at the entire genome and those minor effects. With trait markers, we looked at specific genes,” Arbelaez said. “So when you combine both, you can make better predictions, because now you're accounting for a lot of these different genetic effects.”
And the idea worked. By adding genetic markers and known traits to their model, the team significantly increased its predictive ability, making it 2-10 times more powerful depending on the trait of interest.
Improving small grains
As part of the Small Grains Improvement Lab at U. of I., Arbelaez works on all kinds of crops, specializing in oats as well as rice. He debuted his marker-assisted method in a previous paper, where he and his colleagues evaluated over 500 breeding lines from the lab’s oat breeding program. By using marker-assisted MT-GS for oats, the team improved their model’s predictive ability by around 50%. Since the method translated well to rice, the researchers hope it will be effective for other small grains, too, especially other staple crops.
A rising global population with growing economic power means demand for high-quality grain will continue to increase, so efficient and cost-effective breeding programs are essential. “If you think globally, there are a lot of different market types,” Arbelaez said. “For instance, countries like Peru and Chile tend to like rice that’s slightly stickier than in the rest of the South American countries. So by understanding the alleles that code for texture, we can identify lines of interest for various markets.”
Arbelaez highlighted the importance of international cooperation — he implemented the marker-assisted MT-GS technique while working at the International Rice Research Institute in the Phillipines, and obtained rice samples for this study from the Latin American Fund for Irrigated Rice. Funding for programs like these is essential to feeding a growing population, Arbelaez added. “All agricultural research has been impacted by funding cuts,” he said. “There are revisions happening to a lot of programs that support the work we do.”
Still, Arbelaez is hopeful for the future. “These approaches work across species, and they're extremely useful, especially for these added-value traits. With this research, we proved that we have the tools to tackle small grain quality objectives early on in the breeding process, and that could be extended to other species, such as vegetable crops. So that's great news.”
The study, “Implementing marker covariates and multi-trait genomic selection models to improve grain milling, appearance, cooking, and edible quality in rice (Oryza sativa L.),” is published in The Plant Genome [DOI: 10.1002/tpg2.70068]. Authors include Anup Dhakal, Maribel Cruz, Katherine Loaiza, Juan Cuasquer, Juan Rosas, Eduardo Graterol, and Juan David Arbelaez.
Sources:
Juan Arbelaez, arbelaez@illinois.edu
News writer:
Rami Jameel, rjame2@illinois.edu
Date:
August 7, 2025