What is Breeding, and How Can It Be Made More Efficient?
The rice, vegetables, fruits, and livestock that surround us today are the result of long-term improvements made by human.
This genetic improvement is referred to as “breeding” in technical terms.
In breeding, livestock or plants with useful traits—such as “high yield,” “good taste,” or “disease resistance”—are selected to produce the next generation.
Historically, people have relied on direct observation—seeing and touching—to select superior individuals.
This is known as phenotypic selection.
The term phenotype refers to observable or perceivable characteristics.
For example, traits such as tall stature, large flowers, or sweet fruit.
Specialists in breeding are called breeders.
Outstanding breeders possess years of experience and keen observation skills, and their judgment is sometimes described as “artistic work.”
However, breeding is not an art—it is an essential technology that supports human life.
For example, solving the global food problem requires the rapid and reliable development of superior varieties.
This is where the efficiency of breeding through data-driven approaches has recently attracted attention.
Educational approach
Measuring, Modeling, and Decision-Making: Bringing Science to Breeding with Data
My research focuses on data-driven breeding, which improves breeding efficiency by utilizing genotypic and phenotypic data from plants and animals.
Data-driven breeding can be divided into three main components: “measurement,” “modeling,” and “decision-making”. In our lab, we conduct research that combines data analysis with biological knowledge, such as predictive techniques based on genomic data, high-throughput phenotyping using cameras and drones, and the design of optimal breeding strategies through statistical analysis and simulations.
Through such research, undergraduate students at Komaba campus and those wishing to pursue graduate studies are encouraged to understand the fundamentals of biology, statistics, and informatics, and to take on the challenge of analyzing actual data and constructing models.
Our laboratory emphasizes hands-on skills such as using R and Python, methods for data analysis and hypothesis testing, and experience with data collection in actual fields.
This enables students to cultivate the ability to formulate their own questions and use data to derive answers, and we support their growth into the next generation of researchers who can play an active role in the fields of breeding and life sciences.
Vision for industry-academia collaboration
Advancing Breeding through Data-driven Agricultural Science
In our laboratory, multiple students conduct research on various aspects of “data collection,” “modeling,” and “optimization” related to plant and animal breeding.
Nationwide, there are few laboratories that cover such a broad scope, from collection and analysis to the proposal of optimal breeding strategies.
Our strength lies in our ability to develop and test practical solutions to major challenges in breeding, such as climate change and food security, by harnessing the combined power of data and agricultural science.
Rather than merely collecting phenotypic data, we also integrate it with genomic data to build predictive models, with the ultimate goal of designing optimal strategies that are truly useful in practical breeding.
Looking ahead, we aim to collaborate with breeding-related companies, as well as organizations in the agriculture, food, and environmental sectors, to tackle issues such as sustainable food production, the development of new crop varieties, and problem-solving in the field through AI and data analysis.
We believe that close collaboration between academia and industry will lead to more practical and meaningful outcomes.
