AgTech Takeaways – the expanded menu (part 3): In March, I participated in an Austrade-facilitated Australian Agtech Delegation to the US and wrote a blog about my takeaways with four main points of conclusion. Over the next month each of those key points – integrated systems approaches, farming practice adaptation, genomics / computational breeding and open data – will be expanded on in turn.
3. Plant breeding has become part of the AgTech investment environment. With this sort of investment incentive, traditional plant and animal breeding will be pushed faster than ever before. Genomics and computational breeding are going to have much more potential to produce step change productivity improvements than many of the on-farm applied technology solutions.
The plant breeding industry is undergoing significant disruption which has at its core one common factor: mathematics enhanced by cheaper, more powerful, computing capacity.
Plant breeding has traditionally been a slow and expensive process. Significant infrastructure has been required for field and laboratory trials, and the length of time taken to achieve results has required big budgets and long tenure. This has generally meant that the only organisations with the facilities and capability to carry out plant breeding research have been large private organisations or universities and other public institutions. The field has not been a natural fit for start-ups. Not too many venture capital (VC) funds have been willing to invest in a research station full of field plots and greenhouses – even if the breeding outputs are likely to be significant.
However, this excellent article from Bayer research explains how the breeding process is being shortened significantly by computers and algorithms; two things that sit very comfortably in the start-up world:
Up to now it took six years and around 5000 plants to transfer a characteristic background to a cotton plant. In the simulation, Gene Stacker reduces the number of plants to 1,700 and the time to five years. Thanks to mathematics, a new breeding can then reach the market more quickly. In addition, work and expenses can be considerably reduced – in the case of the experiment with cotton, by 66 percent.
VC funds are now investing in start-up breeding companies whose main capital requirements are computers and lab equipment. Obviously it’s not quite this simple and there are many more steps involved – some of which require significant infrastructure – before finished varieties reach farmers’ fields. Nonetheless, this significant shift in the plant breeding research environment opens this field to start-up companies and VC funding.
This does not mean the traditional players in plant breeding research will become irrelevant – far from it. Computational breeding techniques are speeding up parts of the process, but breeding improvements will always be dependent on scientific discovery and background knowledge to be effective. Knowledge about how plants grow, what traits are important, how shifting climates will change trait behaviour, how diseases interact with plants; all these things remain critical for informing computational breeding about what traits and combinations are desirable.
What VC-backed research will bring to the plant breeding world is urgency. VC funds mature more quickly than traditional plant breeding processes deliver new varieties – which means that whatever they fund will have shorter delivery timelines than what has been the norm. Faster results attract more VC investment – in turn, more VC investment promotes faster delivery.
The pace of that activity will stimulate the existing innovation and research environment as expectations of shortened delivery times are placed on everyone, whether VC or publicly funded.
For example, AI platform FastStack – developed by Professor Ben Hayes at the Queensland Alliance for Agriculture and Food Innovation (QAAFI) with LongReach Plant Breeders – promises to dramatically fast-track the wheat breeding cycle. This GRDC article explains how the platform will track the flow of valuable genes in crop breeding programs and seek out the most likely combinations to improve performance:
Professor Hayes says the problem of long breeding cycles is frustrating given the extraordinary recent advances that have been made in understanding the genetic basis of yield, grain quality and disease resistance traits in wheat. Should FastStack live up to expectations, Professor Hayes hopes to reduce the length of a wheat breeding cycle from 12 years to as few as two or three.
“To improve on the performance of past wheat varieties, breeders must bring together an ever-increasing number of high-performing genes into one wheat genome,” he says. “It is this ‘gene stacking’ process that is so time consuming, but which we can reduce by using the computing power of an AI system.”
Accelerated computational breeding clearly has the potential to enable dramatic productivity improvements, and is a space to keep a close eye on.