Weathering complexity in agriculture

O.R. will be the key to understanding the complex genetics that go into what we eat.

Agriculture

By Dan Dyer

The weather has certainly been on everyone’s mind recently. When you are in farming, the weather is on your mind all the time. Farmers don’t take a break from thinking about it, honestly. That’s because a late spring freeze can kill an emerging crop. If that happens, the replanting cost could reach $100,000 on an average-sized farm. Or, an extended mid-summer dry period can rob 20 percent or more from the yield potential for a crop. For a farmer, that loss can easily be the difference between profit and loss for the entire year.

We all know local weather is a complex system, and our best predictive abilities have wide margins of error. In agriculture, the situation is even more challenging. This is because crop performance is the result of the interaction of two complex systems: environment and genetics. Consider for a moment the complexity of modern medicine. The advances that have been made in understanding disease processes, and then being able to manage disease treatment, prevention and health maintenance, is among the great accomplishments of our lifetime. Much of these advances in diagnostics, biology, molecular biology and biochemistry can be leveraged directly into agriculture.

But, in fact, humans are relatively simple. The genetic diversity in the corn crop alone is more than the genetic diversity among all humans. And that is only one of the crops we work with and certainly not the most complex. The real difference is we humans control the environment we live in. We can adapt our behavior. We put on a coat when it is cold. We have air conditioning when it is too hot. Crops, however, live in the ambient environment. And they have evolved elaborate mechanisms to alter their physiology and biochemistry in response to environmental cues.

Agriculture

Figure 1: Just how important is the impact of environment on agriculture? Figure 1 shows the corn grain production variation over the last five years in North America. Acreage remained roughly constant – the primary source of variation was weather. The production drop in 2012 was valued at $15 billion. One crop, one region, one year with an impact of $15 billion. Chart Source: www.FAO.org/statistics

We are all familiar with the complexity of the environment. We can understand the engines of weather variation such as ocean currents, interacting air masses, solar flux and reflectivity. But the interactions among these components are so complex that the system retains a significant degree of chaos, which we can only deal with in probabilities, even within a single day. In agriculture, we are concerned with the pattern of weather throughout the growing season. Extremes in temperature or rainfall – too hot, too cold, too dry, too wet – cause critical fluctuations in crop yields.

Until recently, we thought the nature of genetics was pretty simple: four nucleotides, arranged in each species in a fixed order, grouped into a fixed number of chromosomes. Changes in the sequence happened occasionally – called mutations – that sometimes changed the characteristic of the individual. These were passed along from each parent to their child. Wonderful, predictable, clear and relatively simple. Boy, were we wrong.

There are four nucleotides. Most species have a fixed number of chromosomes, but not all. In a sugarcane field, you can find individuals that vary widely in the number of chromosomes. Genes don’t act on their own. There are genes that control groups of other genes. There are mutations. But there are also systems that fix them. And, sometimes, those systems cause new mutations. And the genes aren’t all in the same order. In fact, different individuals of the same species don’t always have the same number of genes at all. Some are missing or duplicated or moved to another place on the chromosome, or even a different chromosome. Fascinating and complex.

You can start to imagine the complexity when these two systems, each complex on their own, interact. Until recently, understanding this has been beyond the limits of technology. We didn’t have the necessary data or the computing scale. But now we are entering the age of exploration for agricultural data analytics. The data streams are massive. We have weather data collected at hundreds of locations, with nearly continuous monitoring of temperate, solar flux, wind speed and rainfall. Harvest machines record yields as they progress through the field. We have digital measurements and images from drones and satellites. We produce the molecular genetic fingerprint for hundreds of thousands of new crop varieties every year. It is a data resource that was only a dream a decade ago. And now, we have engaged a wide global community of collaborators to explore novel approaches to deriving the critical insights from this extraordinary data resource. Figure 2 shows a simple example.

Agriculture

Figure 2: In ordinary conditions, this soybean variety is comparable to all others in general. When rainfall is scarce, however, it really stands out. Placing the variety in more drought prone soils and locations will really benefit the grower. How much does this really matter? A five bushel/acre increase in soybean yield in the United States could serve the protein needs of 35 million people. A 5 percent increase in corn yield in the United States could meet the calorie requirements of 50 million people. Given that there will be an additional 5 billion people to feed in 50 years, maximizing the efficiency of production is essential. Source: Syngenta

Being more accurate in matching the genetics to the environment brings enormous social and economic benefits, even in this simple example of data-driven product placement decisions. But this example only touches the surface. Now we are exploring the mechanisms of differential responses at the genetic level, asking such questions as: What are the genes in this soybean variety which enable it to perform better under drought than others? What do those genes do, and how are they controlled?

We all know that different plants flower at different times during the season. Gardeners will recognize that different varieties of the same species also flower at different times. You can arrange a tulip garden, for instance, with varieties that will flower in succession for a long season of blooms. This is actually an important characteristic in agriculture. In most seasons, it gets hotter and drier as the season progresses. Hot, dry weather at flowering can be very detrimental to yield. For vegetable growers, predictable time to flowering is especially important, as they target plantings to specific market periods. But why do different varieties flower differently?

In most plants, temperature triggers flowering – both the accumulated heat during the growth of the plant, and extreme high or low temperatures. For many plants, day length also triggers flowering. But, different varieties respond differently. One variety may be induced to flowering when the day length is less than 14 hours, but for another it may have to be shorter than 13 hours, and for a third, day length may not be a trigger at all. As we begin to understand the molecular biology that controls these responses, we can develop varieties with more stable and predictable flowering, despite environmental variation, and develop varieties better tuned to flower under the most ideal conditions for crop productivity.

Application of operations research and advanced data analytics have already had a profound impact on the efficiency of our breeding and product development programs. Now those same disciplines are being used to build models of how genes and environment interact. This is completely changing the way the crops of the future will be designed and developed, with data-driven insights forming a foundation to feed the future population in an environmentally sustainable way. ORMS

Dan Dyer, head of seeds development at Syngenta, holds a Ph.D. in agronomy and crop science. His Syngenta group develops new crop genetics to promote sustainable agricultural productivity. Throughout his 35-year career, Dyer has been responsible for leading highly dispersed, global innovation teams. Beyond the application of technology to meeting the global need for food security, his interests lie within how the mind works, how ideas are formed and innovation evolves, the nature of effective communication, and why people come to believe what they do.