IN EARLY NOVEMBER 2020, a worrisome e-mail arrived in Roberto Santamaria’s inbox. As the director of health and human services on Nantucket, he closely monitored the weekly reports from Cambridge-based startup Biobot Analytics that relayed how much COVID-19 was swirling in the sewage from the island’s main treatment plant. His eyes always jumped to the graph on the first page: If the line angled up, it was bad news; if it fell, it was good. Today was bad news. Between October 19 and November 2, the amount of virus in the island’s waste water had tripled.
The state had been posting weekly metrics that made Nantucket’s outbreak look well controlled, but daily, local metrics told a different story. Santamaria had carefully noted that Biobot’s predictive estimates of cases on the island had been spot on. In June, sewage data indicated virus levels were spiking. “It is incredibly important that we look at this,” he told the Board of Health at a meeting, “because we know that the sewage reports are indicative of what is to come.”
Up until then, guided by Biobot’s data, the board had taken a series of steps to control the spread of the virus: first a mask order for the island’s busy historic districts — which includes the wood-shingled shops and whaling museum in downtown Nantucket, and the tiny town center of Siasconset — then limits on gatherings, and then special restrictions on construction and landscaping sites after a small cluster of cases in workers. So on that day in November, with the numbers in sewage spiking again, the board voted to reinstate the construction restrictions. “The sewage data informed every step of the way,” Santamaria says now. “It is the single biggest metric we use.”
When Biobot first offered a pro bono service to track coronavirus in hundreds of communities around the country in 2020, few public health officials had ever used the sewage coursing through pipes to measure health. But many of them, like Santamaria, were looking for anything to give them a better handle on the ever-changing pandemic. Biobot’s founders, former MIT researchers Newsha Ghaeli and Mariana Matus, saw an opportunity to make a difference in the pandemic while proving a point they had been trying to make for years: Sewage is a gold mine of health information.
Their move paid off. Many communities, including Nantucket, have signed on to regular waste-water testing services, and Biobot has provided data to more than 700 cities, towns, and institutions in all 50 states, the Virgin Islands, and some places in Canada, Mexico, and elsewhere in Latin America. The company has secured nearly $30 million in venture capital and during the pandemic has grown from five employees to more than 60. Biobot’s work has galvanized media interest, in no small part because of the countless opportunities for poop jokes (a segment on “The Daily Show With Trevor Noah” about the company featured a reporter sitting on a toilet).
But as the pandemic grinds on, it’s clear that far from being a joke, sewage has become something we all need in these uncertain times: a beacon.
Because people shed the virus in their stool before they test positive, Biobot’s data are often a leading indicator of where the infection is heading, arriving ahead of case counts by a few days. “It has many advantages over the current existing data collected by the hospital system,” Matus says. It’s inclusive (even when people fail to get tested or see a doctor, they still poop), it’s dynamic (it takes just a few hours for human waste to travel through the sewer system), and it’s inherently private (everybody’s contribution gets aggregated into the stream).
Those advantages have been clearest in the chart of waste-water levels at Deer Island Treatment Plant in Winthrop. The Massachusetts Water Resources Authority posts it on its public website, updated almost daily, and it’s widely shared on social media. Public health experts closely follow the posts to get a view of the pandemic that isn’t captured by clinical measures such as testing. When Biobot’s analyses were finding record-high levels of COVID-19 late in December, it foreshadowed the magnitude of the Omicron variant-fueled wave that was about to hit the region. And when the spike at Deer Island crested early in January and began to drop, it was heralded in headlines nationwide as a welcome signal that the latest surge might end sooner than feared.
So maybe it’s time to set the poop jokes aside. After all, anyone who’s ever had to give a urine or stool sample to their doctor knows that what we excrete holds valuable information. With sewage, the question has always been: Can we learn something equally valuable from the collective waste of hundreds, or millions of people? Biobot’s founders envisioned sewage as a vital source of information even before the pandemic, and they have an ambitious vision of how it might continue to be used in the future. Matus says they see sewage becoming a “pillar of public health.”
IN AUGUST 2017, Newsha Ghaeli paced the stage at MIT’s Kresge Auditorium in front of a background of illuminated blue and purple triangles. It was MIT delta v’s Demo Day, the culmination of a prestigious three-month “venture accelerator” program designed to foster student entrepreneurs, in which they had a few minutes to pitch their companies to investors and the news media. Ghaeli flashed headlines about recent health-related outbreaks: the opioid epidemic, the recent spread of Zika in children, lead in the drinking water in Flint, Michigan. “Now, what all of these outbreaks have in common is that we only hear about them when crisis turns to catastrophe,” Ghaeli said. Imagine if we had a better way of measuring health before people got sick or died, she challenged her audience. The answer was sewage, and Biobot Analytics was going to use it to create “a human health database powered by everyone in a city.”
Biobot is the product of a project called Underworlds that Ghaeli had managed in collaboration with Mariana Matus for three years. Neither of them set out to become apostles of sewage. Matus came to MIT first, a doctoral student in computational biology from Mexico City who had studied in the Netherlands and the United Kingdom, keen to solve real-world problems at the intersection of biology and the environment. One day, her adviser, human microbiome expert Eric Alm, relayed a conversation he’d had with Carlo Ratti, head of MIT’s Senseable City Lab, speculating on the question: What can you learn about the health of a whole city by looking at its sewage?
“Basically, everybody in the lab just started to laugh,” Matus says. But she immediately wanted to help, and she joined Alm and Ratti in writing a funding proposal that proved successful. By that time, Ghaeli, who was born in Iran and grew up in Canada, had come to Ratti’s lab as a researcher interested in climate change. But when she heard about the project, she says, she was “floored with the concept of looking at sewage as a source of information on human health.”
Waste-water epidemiology wasn’t new, but it was fairly obscure. Sewage had been used to monitor illicit drug use and poliovirus, but it wasn’t clear what other signals of health could be found in the harsh sewer environment and how those signals might be compared across sites within a city or across cities. Ghaeli, Matus, and other faculty and students worked in Boston, Cambridge, Seoul, and Kuwait City to develop new technologies and approaches for collecting sewage from manholes and analyzing it for bacteria, viruses, and chemicals.
Three years into the project, both Ghaeli and Matus were interested in leaving academia for the private sector, and they decided to turn their work into a company. Though neither had business experience, they benefitted from MIT’s deep resources for entrepreneurs. After first pitching their idea at the MIT $100K Entrepreneurship Competition (their project didn’t make it to the final round), they participated in a three-week boot camp, MIT Fuse, and received funding through MITdesignX, a startup accelerator in the architecture department, MIT Sandbox, and the MIT IDEAS Global Challenge. Then they got a spot in delta v, which came with significant funding to spend a summer refining their technology and vision.
Originally, Ghaeli and Matus concentrated on developing sewage-sampling robots — their company’s namesake “biobots” — as well as analysis tools, with the intention of working with city governments to sample sewage from manholes to get a neighborhood-by-neighborhood picture of health. In her Demo Day presentation, Ghaeli focused on the major epidemic on the minds of every policy maker at the time: the opioid crisis. Health officials are usually looking backward at overdose deaths, she said during her presentation, but sewage offers a way to look more proactively at where drug use is happening. Flashing a proud smile, she announced the company had its first paying client, which would make them “Ramen profitable” by the end of the year.
Over the next 12 months, they worked with the town of Cary, North Carolina, which had been experiencing a crush of overdoses, to track opioid use by testing sewage at several sites around the city. Matus says the data helped to inform outreach strategies that ultimately led to a decrease in overdoses. But even with this success, the company still struggled to gain a foothold with policy makers, who were loath to pay anyone to learn about the contents of their sewage.
Then came COVID-19. When reports of its spread first emerged, the company was just Ghaeli, Matus, and three employees. But they realized early on that they needed to take action. “If we want to be a public health company, we have to be responsive,” Ghaeli says. They quickly pivoted, working with Alm’s lab and Nanyang Technological University in Singapore to develop a test to detect and quantify SARS-CoV-2 in waste water. In March 2020, they became the first group to do that in the United States, using samples taken from the Deer Island plant.
They launched a campaign to sample communities across the country on a pro bono basis, setting a goal of 100 sites, which Ghaeli says was “kind of like a moonshot number for us.” They filled the spots in two weeks. And two weeks after that, she says, “we had hundreds of waste-water treatment plants and communities on a waiting list.” They expanded the program to hundreds more clients, began hiring, and formally launched a paid COVID-19 analysis service that June.
“Four years ago, it was our vision and less shared by the collective,” Ghaeli says. But the coronavirus pandemic provided a massive proof of concept for waste water as a public health tool.
Waste-water data is inherently “noisy” — that is, it has a lot of variability from sample to sample that isn’t meaningful — so the information is most useful when there’s a big shift — a signal in the noise. “We went through this long phase of just people getting used to that intrinsic level of noise in the data and losing faith and regaining faith,” Matus says. “Now we’re even seeing some situations where they trust the waste-water data more than the clinical data.” That’s especially true as more people take at-home tests that aren’t reported, and widespread vaccination and milder symptoms from Omicron infections all mean that many infected people are not making their way into the health system.
Watching the rise and fall of COVID-19 in your community’s sewage is more intuitive than parsing case counts and hospitalizations. So even beyond its scientific value, cities are finding waste-water data to be an important communication tool. For example, Chattanooga, Tennessee, participated in Biobot’s initial pro bono program and later contracted with the company for weekly analyses, which it publishes online. State law has hampered the city from instituting mask mandates and other measures, but waste-water data in combination with other metrics have helped officials decide how to deploy resources and when to launch vaccination drives and awareness campaigns. The city’s practice of using Biobot data has garnered news stories and helped bring attention to the state of the pandemic. “Our mayor is using it,” says Mary Lambert, Chattanooga’s head of community health. “It’s something that we can use when we’re trying to talk about vaccination.”
GIVEN THE NUMBER of communities relying on Biobot’s data, the company’s headquarters is surprisingly cramped. It includes a narrow room with computer terminals and a dedicated laboratory about the size of a small restaurant kitchen within a shared floor at The Engine, a startup accelerator in Cambridge’s Central Square and one of the company’s investors. “When I started, there were like ten people,” says Kyle McElroy, Biobot’s principal research scientist; now the company extends into some of the other shared spaces on the floor.
Lab-coated employees mill around carrying tubes of cloudy liquid that are constantly arriving on FedEx trucks (McElroy estimates they process about 200 or so sewage samples each week). The samples are heated to kill any pathogens and filtered to remove what McElroy calls “large particulates” — yes, that’s a euphemism for poop — and then viral particles from the sample are concentrated for analysis. The process is similar to laboratory COVID-19 PCR tests: Chemicals are used to split virus particles open, extract small bits of viral genetic material, and make many copies of the material so the virus can be detected. In some of the samples, Biobot also uses genetic sequencing to identify specific variants of the virus.
Interpreting the results is more challenging than for an individual test, because the contents of sewage vary widely from sample to sample. Biobot calculates a “normalized” virus count based on the fecal concentration of each sample. Its analytics team uses this to create a tailored report for each client that includes an estimate of how many cases per day it can expect to see.
As a company coming of age in the pandemic, Biobot’s physical presence is scattered. In addition to the Cambridge lab, it now has an office in New York City, where Ghaeli has relocated, and many of its new hires (the majority of whom are women) work remotely from other locations. For Matus, the shift to remote work has been one of the bright spots of the pandemic; she herself contracted COVID-19 early in the pandemic, found out she was pregnant two weeks later, and had a baby in November 2020. “Before COVID, I was on a plane every week,” she says, so the ability to continue leading the company at home has been helpful.
Biobot is operating in a very different market than it first did when its founders pitched the concept to investors. Academic research teams around the world are now working on further developing waste-water surveillance technology and approaches. New companies have joined the space. And waste-water epidemiology has become part of federal strategy. Last summer, Biobot received funding from the US Department of Health and Human Services and the Centers for Disease Control and Prevention to launch a 12-week program to collect and analyze data from more than 300 waste-water treatment facilities across the country, representing nearly 30 percent of the US population. The CDC has launched a National Wastewater Surveillance System that is helping to fund and support waste-water data collection nationally, and Biobot’s data will be included.
Amy Kirby, program lead for the surveillance system, says the technology is here to stay: “It is a way to get information on community-level infection trends that is independent of health care access or testing access and reporting.” Even beyond the pandemic, she says, the surveillance system could be used to track antibiotic resistance, foodborne infections, and emerging diseases.
In an increasingly crowded field, Biobot’s leaders plan to focus less on serving as a one-stop shop for laboratory analysis. Instead, they want to innovate in the analytics arena: gleaning new kinds of information from sewage, introducing new technologies for measuring it, and using expertise in epidemiology and data science to interpret data in new ways.
Matus says that the first step is to expand COVID-19 tracking to include multiple variants as well as influenza and RSV, respiratory viruses that have overlapping symptoms and can be hard to disentangle with clinical data alone. Last November, the company announced it had successfully detected influenza virus in waste water. Biobot is running a pilot program with several of its customers this flu season, and continues to offer a product to measure “high-risk substances” like fentanyl, heroin, and other opioids that are highly likely to lead to overdoses.
Ghaeli and Matus still see the potential for using waste water taken from different parts of a city to understand where health problems and outbreaks are happening. But neighborhood-level analysis requires dipping into manholes, a more expensive undertaking than taking the sewage from treatment plants. In the short term, small-scale waste-water analysis may be more useful at the building level; institutions such as universities and large employers are increasingly using waste water to pinpoint outbreaks in dorms and other buildings.
Cambridge is one of the few cities working with Biobot to analyze site-specific coronavirus data on its different neighborhoods, which it publishes on a dashboard. Sam Lipson, director of environmental health at the Cambridge Public Health Department, has collaborated with Ghaeli and Matus on the city’s behalf since their MIT days. “We were interested in this before the pandemic and I’m impressed and amazed at their ability to pivot as quickly as they did,” he says.
He’s hopeful they can continue to track other measures in the city’s waste water.
Ultimately, the larger role of waste-water epidemiology in public health will depend on how reliable it proves in finding other signals — and what governments and institutions are willing to invest in tracking them. Biobot’s founders still believe in the broader vision they pitched before the pandemic. It may be possible to detect complex chemical signatures of disease within the contents of sewage, which could someday allow public health authorities to track noninfectious health problems — cancer, diabetes, obesity, and other serious conditions. “What people say they would love the technology to do is to be able to flag new problems before we even know about them,” Matus says.
Without knowing what to test for, that would require looking for general signatures of health and disease. If scientists can find a way to do that, sewage would serve as a harbinger of epidemics we couldn’t even predict. In the long term, Ghaeli says, “I see us leveraging waste water to essentially keep a pulse on the health of the collective.”
Courtney Humphries is a freelance journalist and a doctoral candidate in environmental sciences at the University of Massachusetts Boston. Send comments to firstname.lastname@example.org.