Artificial Intelligence is already helping to grow some of the food we eat.
The glass-encased cabinet looks like a display meant to hold reasonably priced watches, or drugstore beauty creams shipped from France. But instead of this stagnant merchandise, each of its five shelves is overgrown with leaves — moss-soft pea sprouts, spikes of Lolla rosa lettuces, pale bok choy, dark kale, purple basil or red-veined sorrel or green wisps of dill. The glass structure isn’t a cabinet, but rather a “micro farm.”
The gadget is on display at the Richmond, Virginia headquarters of Babylon Micro-Farms, a company that aims to make indoor farming in the U.S. more accessible and sustainable. Babylon’s soilless hydroponic growing system, which feeds plants via nutrient-enriched water, allows chefs on cruise ships, cafeterias and elsewhere to provide home-grown produce to patrons, just seconds after it’s harvested. Currently, there are over 200 functioning systems, either sold or leased to customers, and more of them are on the way.
The chef-farmers choose from among 45 types of herb and leafy-greens seeds, plop them into grow trays, and a few weeks later they pick and serve. While success is predicated on at least a small amount of these humans’ care, the systems are autonomously surveilled round-the-clock from Babylon’s base of operations. And artificial intelligence is helping to run the show.
Babylon piloted the use of specialized cameras that take pictures in different spectrums to gather some less-obvious visual data about plants’ wellbeing and alert people if something seems off.
Imagine consistently perfect greens and tomatoes and strawberries, grown hyper-locally, using less water, without chemicals or environmental contaminants. This is the hefty promise of controlled environment agriculture (CEA) — basically, indoor farms that can be hydroponic, aeroponic (plant roots are suspended and fed through misting), or aquaponic (where fish play a role in fertilizing vegetables). But whether they grow 4,160 leafy-green servings per year, like one Babylon farm, or millions of servings, like some of the large, centralized facilities starting to supply supermarkets across the U.S., they seek to minimize failure as much as possible.
Babylon’s soilless hydroponic growing system
Courtesy Babylon Micro-Farms
Here, AI is starting to play a pivotal role. CEA growers use it to help “make sense of what’s happening” to the plants in their care, says Scott Lowman, vice president of applied research at the Institute for Advanced Learning and Research (IALR) in Virginia, a state that’s investing heavily in CEA companies. And although these companies say they’re not aiming for a future with zero human employees, AI is certainly poised to take a lot of human farming intervention out of the equation — for better and worse.
Most of these companies are compiling their own data sets to identify anything that might block the success of their systems. Babylon had already integrated sensor data into its farms to measure heat and humidity, the nutrient content of water, and the amount of light plants receive. Last year, they got a National Science Foundation grant that allowed them to pilot the use of specialized cameras that take pictures in different spectrums to gather some less-obvious visual data about plants’ wellbeing and alert people if something seems off. “Will this plant be healthy tomorrow? Are there things…that the human eye can't see that the plant starts expressing?” says Amandeep Ratte, the company’s head of data science. “If our system can say, Hey, this plant is unhealthy, we can reach out to [users] preemptively about what they’re doing wrong, or is there a disease at the farm?” Ratte says. The earlier the better, to avoid crop failures.
Natural light accounts for 70 percent of Greenswell Growers’ energy use on a sunny day.
Courtesy Greenswell Growers
IALR’s Lowman says that other CEA companies are developing their AI systems to account for the different crops they grow — lettuces come in all shapes and sizes, after all, and each has different growing needs than, for example, tomatoes. The ways they run their operations differs also. Babylon is unusual in its decentralized structure. But centralized growing systems with one main location have variabilities, too. AeroFarms, which recently declared bankruptcy but will continue to run its 140,000-square foot vertical operation in Danville, Virginia, is entirely enclosed and reliant on the intense violet glow of grow lights to produce microgreens.
Different companies have different data needs. What data is essential to AeroFarms isn’t quite the same as for Greenswell Growers located in Goochland County, Virginia. Raising four kinds of lettuce in a 77,000-square-foot automated hydroponic greenhouse, the vagaries of naturally available light, which accounts for 70 percent of Greenswell’s energy use on a sunny day, affect operations. Their tech needs to account for “outside weather impacts,” says president Carl Gupton. “What adjustments do we have to make inside of the greenhouse to offset what's going on outside environmentally, to give that plant optimal conditions? When it's 85 percent humidity outside, the system needs to do X, Y and Z to get the conditions that we want inside.”
AI will help identify diseases, as well as when a plant is thirsty or overly hydrated, when it needs more or less calcium, phosphorous, nitrogen.
Nevertheless, every CEA system has the same core needs — consistent yield of high quality crops to keep up year-round supply to customers. Additionally, “Everybody’s got the same set of problems,” Gupton says. Pests may come into a facility with seeds. A disease called pythium, one of the most common in CEA, can damage plant roots. “Then you have root disease pressures that can also come internally — a change in [growing] substrate can change the way the plant performs,” Gupton says.
AI will help identify diseases, as well as when a plant is thirsty or overly hydrated, when it needs more or less calcium, phosphorous, nitrogen. So, while companies amass their own hyper-specific data sets, Lowman foresees a time within the next decade “when there will be some type of [open-source] database that has the most common types of plant stress identified” that growers will be able to tap into. Such databases will “create a community and move the science forward,” says Lowman.
In fact, IALR is working on assembling images for just such a database now. On so-called “smart tables” inside an Institute lab, a team is growing greens and subjects them to various stressors. Then, they’re administering treatments while taking images of every plant every 15 minutes, says Lowman. Some experiments generate 80,000 images; the challenge lies in analyzing and annotating the vast trove of them, marking each one to reflect outcome—for example increasing the phosphate delivery and the plant’s response to it. Eventually, they’ll be fed into AI systems to help them learn.
For all the enthusiasm surrounding this technology, it’s not without downsides. Training just one AI system can emit over 250,000 pounds of carbon dioxide, according to MIT Technology Review. AI could also be used “to enhance environmental benefit for CEA and optimize [its] energy consumption,” says Rozita Dara, a computer science professor at the University of Guelph in Canada, specializing in AI and data governance, “but we first need to collect data to measure [it].”
The chef-farmers can choose from 45 types of herb and leafy-greens seeds.
Courtesy Babylon Micro-Farms
Any system connected to the Internet of Things is also vulnerable to hacking; if CEA grows to the point where “there are many of these similar farms, and you're depending on feeding a population based on those, it would be quite scary,” Dara says. And there are privacy concerns, too, in systems where imaging is happening constantly. It’s partly for this reason, says Babylon’s Ratte, that the company’s in-farm cameras all “face down into the trays, so the only thing [visible] is pictures of plants.”
Tweaks to improve AI for CEA are happening all the time. Greenswell made its first harvest in 2022 and now has annual data points they can use to start making more intelligent choices about how to feed, water, and supply light to plants, says Gupton. Ratte says he’s confident Babylon’s system can already “get our customers reliable harvests. But in terms of how far we have to go, it's a different problem,” he says. For example, if AI could detect whether the farm is mostly empty—meaning the farm’s user hasn’t planted a new crop of greens—it can alert Babylon to check “what's going on with engagement with this user?” Ratte says. “Do they need more training? Did the main person responsible for the farm quit?”
Lowman says more automation is coming, offering greater ability for systems to identify problems and mitigate them on the spot. “We still have to develop datasets that are specific, so you can have a very clear control plan, [because] artificial intelligence is only as smart as what we tell it, and in plant science, there's so much variation,” he says. He believes AI’s next level will be “looking at those first early days of plant growth: when the seed germinates, how fast it germinates, what it looks like when it germinates.” Imaging all that and pairing it with AI, “can be a really powerful tool, for sure.”
Biochemist Longxing Cao is working with colleagues at the University of Washington on promising research to disable infectious coronavirus in a person's nose.
This may sounds like science fiction, but several research groups are already working on such trickster coronavirus drugs, with some candidates close to clinical trials and possibly even becoming available late this year. The teams began working on them when the pandemic arrived, and continued in lockdown.
When the pandemic first hit and the state of California issued a lockdown order on March 16, postdoctoral researchers Anum and Jeff Glasgow found themselves stuck at home with nothing to do. The two scientists who study bioengineering felt that they were well equipped to research molecular ways of disabling coronavirus's cell-penetrating spike protein, but they could no longer come to their labs at the University of California San Francisco.
"We were upset that no one put us in the game," says Anum Glasgow. "We have a lot of experience between us doing these types of projects so we wanted to contribute." But they still had computers so they began modeling the potential virus-disabling proteins in silico using Robetta, special software for designing and modeling protein structures, developed and maintained by University of Washington biochemist David Baker and his lab.
"We saw some imperfections in that lock and key and we created something better. We made a 10 times tighter adhesive."
The SARS-CoV-2 virus that causes Covid-19 uses its surface spike protein to bind on to a specific receptor on human cells called ACE2. Unfortunately for humans, the spike protein's molecular shape fits the ACE2 receptor like a well-cut key, making it very successful at breaking into our cells. But if one could design a molecular ACE2-mimic to "trick" the virus into latching onto it instead, the virus would no longer be able to enter cells. Scientists call such mimics receptor traps or inhibitors, or blockers. "It would block the adhesive part of the virus that binds to human cells," explains Jim Wells, professor of pharmaceutical chemistry at UCSF, whose lab took part in designing the ACE2-receptor mimic, working with the Glasgows and other colleagues.
The idea of disabling infectious or inflammatory agents by tricking them into binding to the targets' molecular look-alikes is something scientists have tried with other diseases. The anti-inflammatory drugs commonly used to treat autoimmune conditions, including rheumatoid arthritis, Crohn's disease and ulcerative colitis, rely on conceptually similar molecular mechanisms. Called TNF blockers, these drugs block the activity of the inflammatory cytokines, molecules that promote inflammation. "One of the biggest selling drugs in the world is a receptor trap," says Jeff Glasgow. "It acts as a receptor decoy. There's a TNF receptor that traps the cytokine."
In the recent past, scientists also pondered a similar look-alike approach to treating urinary tract infections, which are often caused by a pathogenic strain of Escherichia coli. An E. coli bacterium resembles a squid with protruding filaments equipped with proteins that can change their shape to form hooks, used to hang onto specific sugar molecules called ligands, which are present on the surface of the epithelial cells lining the urinary tract.
A recent study found that a sugar-like compound that's structurally similar to that ligand can play a similar trick on the E. Coli. When administered in in sufficient amounts, the compound hooks the bacteria on, which is then excreted out of the body with urine. The "trickster" method had been also tried against the HIV virus, but it wasn't very effective because HIV has a high mutation rate and multiple ways of entering human cells.
But the coronavirus spike protein's shape is more stable. And while it has a strong affinity for the ACE2 receptors, its natural binding to these receptors isn't perfect, which allowed the UCSF researchers to design a mimic with a better grip. "We saw some imperfections in that lock and key and we created something better," says Wells. "We made a 10 times tighter adhesive." The team demonstrated that their traps neutralized SARS-CoV-2 in lab experiments and published their study in the Proceedings of the National Academy of Sciences.
Baker, who is the director of the Institute for Protein Design at the University of Washington, was also devising ACE2 look-alikes with his team. Only unlike the UCSF team, they didn't perfect the virus-receptor lock and key combo, but instead designed their mimics from scratch. Using Robetta, they digitally modeled over two million proteins, zeroed-in on over 100,000 potential candidates and identified a handful with a strong promise of blocking SARS-CoV-2, testing them against the virus in human cells. Their design of the miniprotein inhibitors was published in the journal Science.
Biochemist David Baker, pictured in his lab at the University of Washington.
UW
The concept of the ACE2 receptor mimics is somewhat similar to the antibody plasma, but better, the teams explain. Antibodies don't always coat all of the virus's spike proteins and sometimes don't bind perfectly. By contrast, the ACE2 mimics directly compete with the virus's entry mechanism. ACE2 mimics are also easier and cheaper to make, researchers say.
Antibodies, which are long protein chains, must be grown inside mammalian cells, which is a slow and costly process. As drugs, antibody cocktails must be kept refrigerated. On the contrary, proteins that mimic ACE2 receptors are smaller and can be produced by bacteria easily and inexpensively. Designed to specs, these proteins don't need refrigeration and are easy to store. "We designed them to be very stable," says Baker. "Our computation design tries to come up with the stable proteins that have the desired functions."
That stability may allow the team to create an inhaler drug rather than an intravenous one, which would be another advantage over the antibody plasma, given via an IV. The team envisions people spraying the miniprotein solution into their nose, creating a protecting coating that would disable the inhaled virus. "The infection starts from your respiratory system, from your nose," explains Longxing Cao, the study's co-author—so a nasal spray would be a natural way to administer it. "So that you can have it like a layer, similar to a mask."
As the virus evolves, new variants are arising. But the teams think that their ACE2 protein mimics should work on the new variants too for several reasons. "Since the new SARS-CoV-2 variants still use ACE2 for their cell entry, they will likely still be susceptible to ACE2-based traps," Glasgow says.
Cao explains that their approach should work too because most of the mutations happen outside the ACE2 binding region. Plus, they are building multiple binders that can bind to an array of the coronavirus variants. "Our binder can still bind with most of the variants and we are trying to make one protein that could inhibit all the future escape variants," he says.
Baker and Cao hope that their miniproteins may be available to patients later this year. But besides getting the medicine out to patients, this approach will allow researchers to test the computer-modeled mimics end-to-end with an unprecedented speed. That would give humans a leg up in future pandemics or zoonotic disease outbreaks, which remain an increasingly pressing threat due to human activity and climate change.
"That's what we are focused on right now—understanding what we have learned from this pandemic to improve our design methods," says Baker. "So that we should be able to obtain binders like these very quickly when a new pandemic threat is identified."