Connect with us

Thought Leaders

Can AI Become a Plant Whisperer to Help Feed the World?

With the power of AI and big data, scientists are pursuing exciting new frontiers in decoding the complex world of plant genomes for next-gen custom plant breeding that could revolutionize food security and adaptation to climate change.

A stalk of wheat, a cane of sugar. To most of us, these are merely the raw materials of some of our favorite foods – but for scientists, they represent a complicated puzzle that, once solved, could unlock secrets that could allow us to grow more food with fewer harmful effects on the earth, custom breed new biofuel sources at scale, and help people live longer and healthier lives. Those secrets are locked up in the genome of plants – and with advanced AI tools, scientists are beginning to discover the secrets these genes hold.

AI’s capacity to analyze huge amounts of data opens the door to solving the challenges of better understanding plant genomes. This understanding of the interaction between the genetic elements present in plants and different functionalities can help researchers develop hardier strains of plants, enabling them to better overcome biotic and abiotic stresses such as environmental challenges like changing climate patterns, pest infestation and pesticide resistance.

Plant genomes – even of “simple” plants, like sugarcane – are significantly bigger than human or animal genomes, having evolved over a far more extended period than other forms of life. Plants are polypoidal – where genes or entire genomes are duplicated – and capturing interactions between genes and alleles from various ploidies is a challenge, as some of the ploidies could represent orphan genes of older plant strains that are not necessarily active now.

Researchers aim to identify single nucleotide polymorphisms (common DNA sequences), which they can use to understand how plants function and interact with the environment. Once this is accomplished, researchers can better understand the function of each gene – and use that information to breed plants that can be adapted to human needs. Thus, if researchers wanted to develop a strain of wheat that could be grown in more arid areas, they would attempt to identify genes in wheat that could allow for full growth despite a lack of water.  Not all samples will likely carry this gene, as it could be an orphan and currently dormant gene that was part of a polypoidal genome. Machine learning could analyze the gene and  its interaction with the environment, providing indications of untapped genetic potential for achieving that objective through AI-designed breeding strategies.

While this research could be used to  manipulate plant strains, such genetic engineering is far from the only way for researchers to develop strains of crops that have the desired qualities. Humans have been cross breeding strains of crops for millennia. AI can be helpful here as well – identifying strains for breeding selection that have the highest compatibility and are most likely to yield the desired results.

In addition, AI systems could help predict which method of breeding – hybridization, wide cross breeding, chromosome doubling  – will be the most effective. With in-depth genetic information on plants at hand, researchers can further use machine learning to match up genes with the optimal environments in which they are most likely to thrive. This could result in crops that can endure an extended growing season or the planting of crops in areas that could not sustain them before, thus increasing the food supply for an increasingly populous – and hungry – world. Strains that will be hardier could be developed – more able to resist the ravages of climate change or grow even in areas where urbanization or desertification has set in.

Plant genetic information could also be used to help breed strains of crops that are more resistant to specific pests or diseases. Machine learning could identify the traits of plants that are most appealing to insects or pests – odor, color, etc. – and enable researchers to develop genes that would reduce the appeal of these plants to pests. This could result in reducing pesticide use, developing more environmentally-friendly pesticides designed for specific plants in specific regions, or even individual farms – a type of “personalized agriculture” that is safer, cleaner, and greener.

Before the current capabilities of AI, identifying plant genomes was near-impossible – but now that they have been identified, understanding how they work is impossible without advanced AI technologies like machine learning. With the tools that are now available, researchers will be able to understand plants better, and develop new and better methods to help plants thrive in the face of environmental changes, pollution, urbanization, and other issues that affect plant growth and quality. With advanced machine learning, researchers will be able to unravel the mysteries that plants hold – and use those secrets to create a better future for humanity.

Eyal Ronen is the Executive Vice President of Business Development of Evogene, a computational biology company which has developed a unique computational predictive biology "CPB" platform, which leverages AI and big data for the development of life-science products. Eyal holds a B.Sc and M.Sc. in Agronomy from the Hebrew University of Jerusalem and an MBA from Haifa University.