What if farmers could grow sugarcane in a matter of seconds, not days or weeks? Scientists are doing just that. Of course, these crops are not sprouting from soil. Instead they flourish on a computer screen. Digital plants like these are part of a new movement in agricultural science called “in silico,” where researchers design highly accurate, computer-simulated crops to help speed up selective breeding, in which plants are chosen and replanted to amplify their desirable traits. Scientists believe the future of farming is not just in fields, but in graphics, too. This new area of crop science comes at a precarious time for global food security. The world’s current population is some 7.5 billion people, and the Pew Research Center predicts it will skyrocket to about 9.6 billion by 2050. To make matters worse, researchers have recorded severe drop-offs in soil nutrients and water availability worldwide. As the foundation of how we feed ourselves, future crops will need to make more with less. The millennia-old strategy of just handpicking and replanting the varieties that thrive is too slow, says Eberhard Voit, a biologist at Georgia Institute of Technology. “We need a more targeted approach,” he says. This is where crops in silico may help. By studying plant growth using computer simulations, researchers could discover which attributes make the best pickings and why, in far less time than a traditional growing season. The term in silico, or “in silicon”—refers to silicon computer chips. The technique begins with scientists collecting data about plant behavior under microscopes and in the field. Next they build statistical models that identify mathematical relationships in the data. Researchers then create simulations based on those equations, which allows them to see the traits they measured play out on a screen. Once they visualize the crops, scientists can manipulate the data to see which factors result in the fastest-growing, most drought-resistant or least pest-susceptible plants possible. The digital sugarcane, described earlier this year by researchers at the University of Illinois at Urbana–Champaign, illustrates how crops in silicomight aid farmers. In the graphic, a leafy canopy mirrors the range of heights, leaf sizes and angles measured in real Brazilian sugarcane fields. Sunlight filters through the foliage with varying intensity, missing patches shaded by other leaves. In response to these varied shapes and light exposures the digital crop mimics how real-world sugarcane plants would grow.
Long, Zhu and their team watch a day’s worth of sunlight exposure on symmetrically planted sugarcane. The hotter the color, the more photosynthesis that is occurring. Credit: Wang, Y., Song, Q., Jaiswal, D. et al. Bioenerg. Res. (2017) 10: 626. https://doi.org/10.1007/s12155-017-9823-x
In their study the researchers used the simulation to test four different planting patterns—a perfectly symmetrical grid of seed lines versus a staggered one, and north-to-south versus east-to-west orientation of planting rows. Their model revealed asymmetrical and north-to-south alignment produced the highest yield, which was 10 percent greater than that which typical Brazilian sugarcane fields currently supply. Supercomputers took an entire day to process the team’s data and provide results, says study co-leader Stephen Long at Illinois, who is also a professor of plant biology and crop science at Lancaster University in England. Still, a day of processing is hardly any time compared with an entire growing season. But Long thinks it will not be long before technological advances speed up these calculations. “Within 12 months, what took a day could be done in a minute,” he says. The researchers’ model helps reveal how conditions such as sunlight and shading affect crop growth—but many other factors such as water availability and microbial interactions determine the length of a corncob or width of a soybean. Before these other elements get translated into code scientists first need to understand how they work in real life. Plant physiologists and biologists around the world are now investigating these crucial questions in both the field and lab. For example, Long and Illinois colleague Xinguang Zhu, who also holds a position at the Chinese Academy of Sciences, study a type of photosynthesis unique to crops like soybeans and rice. Voit researches what controls the production of a plant wall–toughening compound called lignin. Jonathan Lynch, a plant physiologist at The Pennsylvania State University and another founder of the crops in silico movement, looks at root behavior under a range of soil nutrient conditions. As they learn more, scientists will factor these key details of crops’ survival and growth into their labs’ in silicosimulations. The next hurdle becomes how to get these different simulations to talk to one another, says Amy Marshall-Colón, a plant systems biologist at Illinois who leads a lab group independent of Long’s and Zhu’s. Research teams conduct their individual research in whatever software they prefer, and many of their programs cannot run simultaneously and produce a plant that makes sense. This is where the National Center for Supercomputing Applications (NCSA) comes in. Housed at Illinois, this computational facility has helped other scientists, such as astrophysicists, translate their math into animations. Programmers at the NCSA are building a software framework that can combine all the individual crop models into one plant displaying multiple programmable features. Using this multitasking framework, researchers can run any assortment of simulated plant features they would like. “You can plug and play in order to answer a specific biological question,” Marshall-Colón says. To achieve this later phase, the burgeoning in silico field will need to grow substantially, especially in the U.S. Lynch, who also has a position with the University of Nottingham in England, feels Europe is more willing to invest in crop modeling. “We’re just not even playing the game” in the U.S., he says. Still, American researchers have found some funding via their universities as well as organizations like the Foundation for Food and Agriculture Research—a nonprofit meant to help the U.S. Department of Agriculture find and promote innovative food production techniques. Scientists working on crops in silico anticipate the growing potential of this young field, and what they will learn about foods essential to human survival. “It is going to become more important in years ahead,” Lynch says, “This is the frontier [of] biology.”
Digital plants like these are part of a new movement in agricultural science called “in silico,” where researchers design highly accurate, computer-simulated crops to help speed up selective breeding, in which plants are chosen and replanted to amplify their desirable traits. Scientists believe the future of farming is not just in fields, but in graphics, too.
This new area of crop science comes at a precarious time for global food security. The world’s current population is some 7.5 billion people, and the Pew Research Center predicts it will skyrocket to about 9.6 billion by 2050. To make matters worse, researchers have recorded severe drop-offs in soil nutrients and water availability worldwide. As the foundation of how we feed ourselves, future crops will need to make more with less. The millennia-old strategy of just handpicking and replanting the varieties that thrive is too slow, says Eberhard Voit, a biologist at Georgia Institute of Technology. “We need a more targeted approach,” he says. This is where crops in silico may help. By studying plant growth using computer simulations, researchers could discover which attributes make the best pickings and why, in far less time than a traditional growing season.
The term in silico, or “in silicon”—refers to silicon computer chips. The technique begins with scientists collecting data about plant behavior under microscopes and in the field. Next they build statistical models that identify mathematical relationships in the data. Researchers then create simulations based on those equations, which allows them to see the traits they measured play out on a screen. Once they visualize the crops, scientists can manipulate the data to see which factors result in the fastest-growing, most drought-resistant or least pest-susceptible plants possible.
The digital sugarcane, described earlier this year by researchers at the University of Illinois at Urbana–Champaign, illustrates how crops in silicomight aid farmers. In the graphic, a leafy canopy mirrors the range of heights, leaf sizes and angles measured in real Brazilian sugarcane fields. Sunlight filters through the foliage with varying intensity, missing patches shaded by other leaves. In response to these varied shapes and light exposures the digital crop mimics how real-world sugarcane plants would grow.
In their study the researchers used the simulation to test four different planting patterns—a perfectly symmetrical grid of seed lines versus a staggered one, and north-to-south versus east-to-west orientation of planting rows. Their model revealed asymmetrical and north-to-south alignment produced the highest yield, which was 10 percent greater than that which typical Brazilian sugarcane fields currently supply. Supercomputers took an entire day to process the team’s data and provide results, says study co-leader Stephen Long at Illinois, who is also a professor of plant biology and crop science at Lancaster University in England. Still, a day of processing is hardly any time compared with an entire growing season. But Long thinks it will not be long before technological advances speed up these calculations. “Within 12 months, what took a day could be done in a minute,” he says.
The researchers’ model helps reveal how conditions such as sunlight and shading affect crop growth—but many other factors such as water availability and microbial interactions determine the length of a corncob or width of a soybean. Before these other elements get translated into code scientists first need to understand how they work in real life.
Plant physiologists and biologists around the world are now investigating these crucial questions in both the field and lab. For example, Long and Illinois colleague Xinguang Zhu, who also holds a position at the Chinese Academy of Sciences, study a type of photosynthesis unique to crops like soybeans and rice. Voit researches what controls the production of a plant wall–toughening compound called lignin. Jonathan Lynch, a plant physiologist at The Pennsylvania State University and another founder of the crops in silico movement, looks at root behavior under a range of soil nutrient conditions. As they learn more, scientists will factor these key details of crops’ survival and growth into their labs’ in silicosimulations.
The next hurdle becomes how to get these different simulations to talk to one another, says Amy Marshall-Colón, a plant systems biologist at Illinois who leads a lab group independent of Long’s and Zhu’s. Research teams conduct their individual research in whatever software they prefer, and many of their programs cannot run simultaneously and produce a plant that makes sense. This is where the National Center for Supercomputing Applications (NCSA) comes in. Housed at Illinois, this computational facility has helped other scientists, such as astrophysicists, translate their math into animations. Programmers at the NCSA are building a software framework that can combine all the individual crop models into one plant displaying multiple programmable features. Using this multitasking framework, researchers can run any assortment of simulated plant features they would like. “You can plug and play in order to answer a specific biological question,” Marshall-Colón says.
To achieve this later phase, the burgeoning in silico field will need to grow substantially, especially in the U.S. Lynch, who also has a position with the University of Nottingham in England, feels Europe is more willing to invest in crop modeling. “We’re just not even playing the game” in the U.S., he says. Still, American researchers have found some funding via their universities as well as organizations like the Foundation for Food and Agriculture Research—a nonprofit meant to help the U.S. Department of Agriculture find and promote innovative food production techniques.
Scientists working on crops in silico anticipate the growing potential of this young field, and what they will learn about foods essential to human survival. “It is going to become more important in years ahead,” Lynch says, “This is the frontier [of] biology.”
