Forget million-dollar labs—Valerio Hoyos-Villegas of McGill University is building tools that work for everyone.
In East Lansing, Michigan, a quiet revolution is taking shape—not in the fields, but just above them. Drones hum gently overhead. Sensors flicker with activity. And deep in the data, patterns emerge—patterns that might just reshape how humanity feeds itself.
At the heart of this shift is Dr. Valerio Hoyos-Villegas, assistant professor at Michigan State University, adjunct professor at McGill University, and former president of the North American Plant Phenotyping Network (NAPPN). His mission? To bring the power of phenomics—the science of measuring traits—to every breeding program, no matter its size or budget.
“Breeding is a resource management game,” Hoyos-Villegas says. “Phenomics gives us the precision tools to manage those resources better—and to unlock genetic gains we’ve only dreamed of.”
The Bottleneck No One Talks About
To the uninitiated, plant breeding might seem like a simple act: select the best plants and grow them again. But in the modern world, it’s a high-tech race to develop crops that resist drought, thrive in poor soils, and yield more with fewer inputs. And while the sequencing of plant genomes has exploded, offering breeders more genetic markers than ever before, one stubborn bottleneck remains: phenotyping.
Phenomics—the systematic measurement of traits—is the bridge between genotype and reality. Without it, genetic predictions are just educated guesses. Yet gathering this data is time-consuming, expensive, and often inaccessible to smaller programs.
Hoyos-Villegas and his collaborators set out to change that.
A New Kind of Toolkit
The team began by asking a deceptively simple question: what do breeders mean when they talk about affordable phenomics?
By analyzing over 100 scientific papers that used the phrase, they created a word cloud of recurring terms. Some patterns jumped out: breeding, genetics, UAVs (unmanned aerial vehicles), yield, and drought. Clearly, affordable phenomics was seen as a tool to improve selection for high-yielding, drought-resistant crops—precisely the traits critical in an era of climate instability.
But affordability, as Hoyos-Villegas points out, is relative. What’s affordable to one lab may be completely out of reach for another. The challenge becomes not just building tools, but scaling them—and designing them to meet breeders where they are.
The UAV Advantage
One of the most accessible phenomics tools today is UAV-based imaging. Small drones outfitted with RGB or multispectral cameras can fly over breeding plots, capturing thousands of data points in minutes. For example, a drone flying at 20 meters altitude can gather high-resolution images that help track canopy temperature—a proxy for water use efficiency—or monitor plant height and growth rate over time.
Incorporating these tools into breeding programs isn’t just about convenience. It’s about increasing selection pressure—that is, the ability to choose only the very best plants. In one ryegrass breeding program in New Zealand, researchers showed that pushing selection intensity up from 5% to 20% could double genetic gain. But the cost was astronomical—unless phenomics was used. With LiDAR imaging added to the pipeline, the cost was cut in half.
“Increasing selection intensity is the key to faster progress,” Hoyos-Villegas explains. “Affordable phenomics makes that possible.”
Trait Dissection: The Power of Pieces
Beyond just measuring traits, phenomics enables trait dissection: breaking complex traits like yield or drought tolerance into their genetic components. This matters because not all traits are created equal. Some have high heritability and are easy to measure; others are messy, context-dependent, and genetically complex.
A critical insight from Hoyos-Villegas’ work is that increasing the number of individuals in a breeding panel—not just the number of genetic markers—can reveal far more trait-associated loci in genome-wide association studies (GWAS). In other words: more data on more plants, even if it’s cheaper and lower resolution, may be more powerful than piling on high-tech gear for a few elite lines.
The Sensor Fusion Frontier
One emerging direction is sensor fusion: combining data from multiple sensors—like RGB cameras, thermal imaging, LiDAR, and hyperspectral scanners—to build a richer portrait of each plant.
But here too, cost and complexity loom large.
“Custom sensor pipelines offer precision,” Hoyos-Villegas notes, “but they also require intense collaboration between breeders, engineers, and data scientists. Commercial solutions are easier to deploy, but may sacrifice nuance.”
The key is to design systems that scale—solutions that can work in a soybean field in Iowa, a rice paddock in Bangladesh, or a smallholder maize plot in Kenya.
From Data to Decisions
Ultimately, phenomics is not just about collecting more data—it’s about making better decisions.
Does this tool help move the average of the population in the right direction? Does it reduce costs and increase speed without sacrificing accuracy? These are the questions Hoyos-Villegas believes should guide adoption.
One warning he offers: as phenomics tools become more automated, there’s a risk of oversimplifying breeder intuition. A sensor might identify a plant with the “ideal” canopy temperature, but it won’t know if that plant is late-maturing or prone to lodging. Breeders still need to balance complexity with the clean lines of algorithmic selection.
So what does “affordable” really mean?
Hoyos-Villegas suggests that affordability must be defined not by a price tag, but by outcomes. When does the cost of phenotyping outweigh the value of the genetic gain it enables? When does a pipeline stop being a tool and become a barrier?
The answers will differ from place to place—but the goal remains the same: build breeding systems that are more just, more global, and more effective.
A Revolution From Below—and Above
In the end, the phenomics movement is less about technology than it is about equity. The dream is not simply faster or fancier data—it’s about putting power in the hands of more breeders, in more parts of the world, to select for the crops their communities need most.
“We’re not just chasing yield,” Hoyos-Villegas says. “We’re chasing access. And that means building tools that work for everyone.”
As drones lift off from the test plots at Michigan State and data streams into the cloud, the next generation of plant breeding is taking shape. Quietly. Incrementally. But irreversibly.
And this time, it might just change the world from the air down.
