Plant breeders have never had more technology at their disposal. Genotyping, increasingly sophisticated data tools and now AI is changing what breeding programs can see, predict and act on. More capability also brings harder questions about what to adopt, where to put trust and how breeders themselves can add the most value.
That’s why Seed World reached out to Dagan Mor at Gene-G. The Israeli company works alongside breeding programs to apply genetic and analytical technologies, giving its team a close-up view of how breeders are navigating this rapidly changing toolbox. We asked Gene-G to look beyond the technology itself and share what it is seeing: what’s genuinely changing, where the risks are emerging and what the strongest breeding programs may do differently next.
What are breeding programs doing with genetic data today that simply wasn’t practical five years ago?
For years, genotyping outpaced breeders’ ability to use it. Like books before printing, the data was readable only by bioinformatics teams that only large seed companies could afford. For smaller programs, even a 36-SNP purity test was hard to read and use on a daily base, let alone a 120-SNP MABC panel. With AI and analysis-focused genotyping services, genotyping has become what printed books became: an everyday tool for MAS and MABC in programs of every size.
Where do you see AI genuinely changing plant breeding today, rather than simply generating excitement?
The most concrete change we see is the significant progress of genomic-based prediction tools within programs outside of major field crops like corn and soybean. Vegetable and flower breeding programs that cannot afford nor manage the required large trials of phenotyping and genotyping are increasingly utilizing AI-based systems where genomic data from an external service provider like Gene-G supports breeders in making their choices. For example, one of our customers is building an AI tool to predict what a cross will produce before it’s ever made. Instead of making hundreds of crosses and waiting seasons to see results, breeders using these tools and genomic information can now focus on the most promising combinations, making their selections in silico (using simulations and computational models). Tools like this run on genotyping data, which makes good, well-analyzed data more valuable than ever.
As predictive tools improve, what part of a breeder’s judgement becomes more valuable?
As prediction takes over more of the “how”, the breeder’s most valuable judgment is the “what”: which traits are worth breeding for, which problems genetics can actually solve, and what growers will want to grow and consumers will want to eat years from now.
Beyond the traditional targets of yield and disease resistance, we’re seeing breeders take on harder, more complex traits like flavor, texture and fruit appearance. That’s good news for everyone.
What separates breeding programs that use new technology well from those that simply accumulate more tools?
If all you have is a hammer, everything looks like a nail. With so many genotyping and analysis options available, the risk is letting the tools set the agenda. The strongest programs start with the breeding question and choose the tool that fits it. A good breeding program sets its goal on the vision of the breeders/company, on the future needs of their customer, and at any part of the supply chain. from the grower’s field to the dinner table of the consumer. In a backcrossing program, for example, dedicated marker-assisted workflows, like the ones we run with breeding partners, allow breeders to recover the recurrent parent with the desired additional trait in fewer generations, cutting costs and bringing varieties to market faster.
Are breeders becoming better at knowing which material to eliminate early, or are we creating new risks by making those decisions sooner?
Breeding used to and in some parts is still seen as an art. Breeders base their decisions and choices on their experience and their intuitive understanding of the crop, and heavily rely on large numbers. Their main goal is avoiding two mistakes: keeping the wrong plants or discarding the more valued ones. Analyzed genotyping data across an entire population reduces both, which means breeders can have a clearer view on their program, allowing them to make better decisions, eliminating the plants they do not need and moving faster towards their goals.
Where is technology creating new problems for breeders rather than solving them?
Technology itself isn’t the problem; how we use it can be. We have email, messaging and video calls, yet a face-to-face meeting is often still the best conversation. Plant breeding is similar. The new risk is breeders can spend more time with data than with plants. Genetic data will never be the whole answer, and I hope breeders keep walking the fields and tasting the plants we’re all going to eat. Technology is a strong engine adjusting to the breeding cart, but a driver’s firm hand and clear direction are definitely required
What will the best breeding programs be doing differently five years from now?
Breeders who once relied only on instinct and field observation are now adding objective phenotypic measurements and genetic data to their decisions. Five years from now, the best programs won’t be chosen between the two. They’ll use genotyping routinely, and will also understand the genomic structure of their germplasm, predict crosses before making them and screen early with data. This will allow their breeders to focus on what only they can do: setting the goals and making the final calls.
