When farmers want crops that have a better shot at resisting disease or taking up nutrients, they use breeding techniques to develop new varieties. Crossing plants to create these varieties is an agricultural practice that dates back thousands of years.
Breeding has since become more sophisticated, incorporating sophisticated screening technologies to find desirable traits. But the expense of these technologies puts them beyond the reach of farmers and small companies. Benson Hill Biosystems, which splits its operations between St. Louis, MO, and Durham, NC, is trying to level the field for those who don’t have access to advanced breeding technology. The agtech startup’s software, called CropOS, predicts biological outcomes, which helps farmers, greenhouses, and researchers discover traits and make better breeding decisions, says CEO Matthew Crisp.
CropOS works by analyzing vast amounts of genomic and biological data from both public and private sources. The software can then pinpoint which plants will produce the desired traits and improved performance that ag companies and growers want. Everything CropOS learns from analyzing data becomes part of a pool of information that grows as the software’s users add their data to the system. Crisp calls it “a networked model.”
There’s precedent for applying big data analysis to life sciences discovery, notably in the search for new drugs. But this approach is also taking hold in agtech research. Indigo, a Boston-based startup whose microbial seed coating was developed to help cotton plants hold up better in drought conditions, uses analytics to discover and develop its products. After the company raised $100 million in a Series C round in July, CEO David Perry explained to Xconomy that Indigo’s software uses machine-learning techniques to predict which microbes will help a plant when confronted with a stressor, such as drought.
Syngenta (NYSE: SYT) applies analytical tools to soybeans to make the process of designing, developing, and testing new varieties more efficient. The company says these analytical tools ultimately lead to soy varieties with higher yields while also reducing the cost of such research. Both Indigo and Syngenta crunch big data to develop their own products. Davis, CA-based Arcadia Biosciences (NASDAQ: RKDA) uses its technology to develop traits, such as nitrogen use efficiency. Arcadia then licenses those traits to seed and agbio companies that bring those traits into crops such as wheat, soy, and rice. That’s similar to Benson Hill’s approach. But the business model of Crisp’s company takes a page from the tech sector.
Cloud computing made storage and analysis of vast amounts of data accessible and affordable to small companies, democratizing information technology. Benson Hill is trying to do the same thing for smaller players in the ag sector. Crisp calls the approach “cloud biology.” But rather than providing data storage, Benson Hill offers software tools that help companies analyze big data in plants.
CropOS is the cognitive engine that drives Benson Hill’s technology. On top of that platform, the company offers an application called Reveal, which is used to discover and design trait candidates. Another application, Breed, helps make faster breeding decisions. Benson Hill markets its technology to businesses and farmers looking for ways to boost their own plant research efforts. Customers pay a subscription fee to access the software, which allows them to upload their crop data into the system. Analyzing that data, the software helps customers identify which seeds will produce the desired traits, allowing researchers to bypass what would ordinarily take multiple generations of experimentation, Crisp says.
Companies keep ownership of their data and one company can’t see another’s competitive work. But each new piece of biological information that CropOS uncovers becomes part of a blinded pool of data. The machine-learning capabilities of CropOS help the system improve and make better decisions over time, making the system more powerful and more predictive than the efforts of any one company working alone, Crisp explains.
“I like to refer to it as a community,” Crisp says. “Everyone shares the collective benefit of that platform.”
Benson Hill’s technology came out of research at the Donald Danforth Plant Science Center in St. Louis. Benson Hill licensed the technology from the center and formed in 2012. The company is named for Andrew Benson and Robin Hill, two scientists known for their contributions to photosynthesis research. Photosynthesis was Benson Hill’s initial focus, and the company’s technology can screen for the traits that improve the efficiency of that process.
When Benson Hill started, the company used CropOS for its own discovery efforts. The company still has its own internal discovery program. But as a small startup, Benson Hill does not have the resources to field test its discoveries. Instead, the company licenses traits to others. Agribusiness companies developing products from Benson-Hill-discovered traits include potato giant Simplot and Brazilian sugarcane researcher CTC. Benson Hill draws revenue from upfront payments, as well as additional fees to support ongoing research. These deals are also structured to pay Benson Hill royalties on products produced from successfully commercialized traits.
The pool of information feeding CropOS is getting bigger. Benson Hill recently signed on Beck’s Hybrids as a customer. The Atlanta, IN-based seeds company provides seeds to farmers throughout the Midwest. Beck’s plans to add CropOS to its own breeding tools with the goal of improving traits such as plant health, maturity, and yield. The Beck’s partnership followed an alliance with the National Corn Growers Association (NGCA), the corn industry’s largest organization with a membership that includes more than 40,000 farmers. Other small innovators gaining access to CropOS through the NCGA partnership include academic labs, small seed companies, and even some food companies that operate their own breeding operations.
Interest in big data tools in agriculture is evident in the growing investment in the space. “Decision support technology,” a category that includes big data software, accounted for $295 million invested in 46 deals in 2015, according to online agtech investment marketplace AgFunder. One of those deals was Benson Hill’s $7.3 million Series A round. The company is now seeking $15 million in a Series B round, which Crisp says will be used to further develop CropOS, as well as pursue applications of the technology in gene editing. Benson Hill is also looking for more partners to feed the data pool for CropOS.
“We’re really agnostic to the crop or the target,” Crisp says. “Our job is to build a robust tool that can be used by a large body of innovators.”