Advanced Analytics Leads to Big Improvements in Sampling Wastewater for SARS-CoV-2

Researchers and Innovative Technology Companies are Already Making Big Improvements Toward Scaling This Breakthrough Approach to Community Testing

By Carol Brzozowski
September 14, 2020

In Notre Dame University’s Environmental Change Initiative laboratory, public health engineer and post-doctoral research associate Aaron Bivins, Ph.D., P.E., gazes at a computer screen, analyzing wastewater sample data for markers of the presence of SARS-CoV-2 (the virus that causes COVID-19).

“Monitoring wastewater shows tremendous potential as a tool to monitor COVID-19 trends within a community,” says Bivins, an environmental microbiologist. “The aggregated signal of domestic wastewater provides another channel of information that is independent of clinical testing. “Given that a single test of wastewater yields information about an entire community whereas a clinical test only yields information about an individual, the economies of scale are also promising.”

Some utilities have expressed interest in Wastewater Based Epidemiology (WBE) and are intrinsically motivated to lead the way, says Bivins. “They already had the personnel and capabilities on hand when the pandemic began, so they had been able to respond in a timely manner,” he adds.   

Case in point: Tempe, Arizona and Arizona State University have partnered to study the city’s wastewater using WBE for the presence of opioids, cocaine, alcohol, marijuana and now the SARS-CoV-2 virus.

Work such as that done by Bivins and his team is being replicated in about a dozen other universities as researchers seek to create best practices and standard operating procedures for analyzing and normalizing the data. “We want to help with the current pandemic,” says Bivins. “We’re also interested from an academic perspective in understanding and developing the field.  Our current approach is to analyze the wastewater, generate the data, and report our data back to the municipality for them to decide how they want to use it.”

WBE dates to the 1940s, when the wastewater was monitored for the polio virus, notes Bivins. “Scientists would collect sewage samples and inoculate lab animals to see if those animals developed polio virus, which is a very crude approach,” he adds. “From there, it developed into monitoring drug use patterns. There are very important differences between monitoring a chemical signal and a biological signal.”

While sampling wastewater for viruses is not new, it has gained a great deal of attention with
the arrival of SARS-CoV-2

Sampling human waste in wastewater for SARS-CoV-2 is one more tool in the toolbox of indicators and tracing that “gets way ahead in detection because people shed this in their waste well before they exhibit any symptoms.” notes Kevin Marsh. Marsh is the president of FlowWorks, Inc., a SaaS system that collects, manages and analyzes environmental data for municipal and engineering clients. FlowWorks uses a specially-designed cloud-based mass storage service that collects and stores extremely large data sets from virtually any source. Collected data is available immediately through the FlowWorks application and/or its API to effectively monitor, model and predict events. The software suite pre-aggregates information from any sensor or source.

“FlowWorks is in a unique position because we’ve recently created a machine learning and advanced analytics algorithm that, with the right data inputs, calculates in near-real time the infiltration and inflow (I&I) coming into a collection system,” says Marsh. “When we created this particular tool for engineers to quickly and automatically quantify the amount of unwanted rain water entering the sewer system (I&I), COVID-19 didn’t even exist. But it turns out it can play an important role in analyzing wastewater for SARS-CoV-2. There’s monitoring, measuring, and lab analysis but somebody has to make sense of that data. We already have algorithms in place that can be adjusted quite nicely to help better understand the prevalence of something like SARS-CoV-2 in a community.”

FlowWorks data often includes flow rate and temperature – factors that affect raw sampling data for SARS-CoV-2. In addition to normalizing the raw data for those factors, the I&I from rainwater coming into the collection system is another factor to consider as it artificially dilutes samples. In communities lacking flow monitors, FlowWorks has tools to calculate flow rates based on pumping activity at sewage lift stations. All of this data, including calculated I&I, is available on the FlowWorks platform or can be pushed as a live data source to desktop applications like Power BI and Microsoft Excel. Toward the goal of speeding the creation of a robust and reliable SARS-CoV-2 testing solution, FlowWorks has waived its platform fees for governments, utilities, consultants and commercial sampling and analysis enterprises seeking to perfect and scale its use.  

The University of Notre Dame campus in South Bend, Indiana.

The Notre Dame research team is working on two WBE projects: one on its campus and another in a nearby community that had a large COVID-19 outbreak and for which FlowWorks is being used to refine techniques and determine feasibility. Bivins points out that “microbiologists tend to work with one particular method and get quite good at that method. That’s usually their preferred approach because they have a lot of experience with it. “It’s no different in this instance,” he adds. “Lots of research groups are using different approaches. We’re trying to get everyone to be transparent about the method they’re using, the caveats, and the limitations so we can make an informed decision about a standard method.”

Notre Dame researchers are using FlowWorks to “log onto RNA data as well as the flow rate and the influent temperature – different variables that may end up being significant for interpreting this wastewater data,” says Bivins. Bivins and his team began collecting samples on June 8, sampling daily for two months, then sampling three times a week and by summer’s end had three months of data.

There have been a few challenges.

“A lot of the microbiology world is responding to this pandemic with increased testing of the clinical samples and now wastewater samples,” says Bivins. “A lot of the reagents and materials we use are back-ordered. That’s an ongoing issue. “We had samples frozen for a bit. In the case of this particular municipality, we did the first step of the process before we froze them, which helps keep them more stable.”

Several rain events occurred during the sampling of the nearby municipality’s wastewater. “I’m using FlowWorks to experiment with different modes of adjustment,” says Bivins. “I’ve not reached any firm conclusions on the best way of doing that, but one of the big things about FlowWorks is the ability to adjust for inflow and infiltration from rainfall, which makes a big difference in our flow rate.”

“One of the things about this particular virus is that the RNA is present at a low enough concentration,” says Bivins. “In order to detect it on our instruments, we need to take 100 milliliters of wastewater and filter it as such that we trap all the viral particles. That’s known as concentration. “The next step is extraction. We put the filter through a chemical process that pulls the RNA out of the viral particles. At the end of the process, we’re left with purified RNA.  The third step is RT-ddPCR to amplify this RNA in order to get a signal that you can analyze.” The process takes the RNA from a few copies to millions of copies such as the lab can detect it and quantify how much was in the original sample, says Bivins.

While it’s “very much open-ended” as to the variables most important to control for, flow rate is significant, says Bivins. “A wastewater collection system is very complex, hydraulically,” he says. “The flow rate increases and decreases throughout the day depending on whether it’s a weekend or a weekday. We know we need to adjust for the flow that’s coming to the treatment plan each day. Rainfall and infiltration can increase the flow rate while decreasing the RNA concentration.”

Normalizing raw sample data for I&I, flow rate, temperature, and potentially other variables
should produce more consistent and actionable data.

Initially, team members thought temperature would be an important factor. “The theory was this RNA would probably degrade faster when it’s warmer and that is the case,” says Bivins. “But the preliminary data indicates that this RNA is actually fairly stable and robust in wastewater.”

Another component being monitored is fecal content, given that wastewater includes both domestic and industrial inputs.

The team also is measuring in each sample the pepper mild mottle virus – a virus that infects peppers and is in human feces but is not infectious to human blood – to normalize the virus RNA to the fecal content of the sample, says Bivins.

One of the biggest challenges to getting meaningful WBE data for the SARS-CoV-2 is the fact that wastewater is a difficult matrix to work with for microbiologists, notes Bivins.

“Hydraulically, the system changes quite rapidly,” he says. “Wastewater systems collect all kinds of chemicals that affect the reactions. We actually get pretty low recoveries – the recovery is the percent of this virus that from start to finish comes through the entire process.

“For every hundred viral particles, you’re only going to catch some percentage of those all the way to the very last step because there’s inefficiency. The recovery efficiencies right now for our process are around five to eight percent. We’re working on trying to improve that recovery.”

Different research projects are using different methods, which is part of why Notre Dame runs process control, says Bivins. “When we collect the sample, we take another virus and intentionally add it to the sample and then we measure that virus at the end, he says. “By doing that, we can compare that we added this much and we got this much out. This is our recovery percentage.”

By running the correct controls, the team can make some comparisons about methods and interpret the data generated by those methods in light of that information, Bivins added.

“The key with this wastewater data is that most of the difficulty is in the interpretation. For our lab at Notre Dame, taking a wastewater sample and measuring the SARS-CoV-2 has become fairly routine,” he says. “But knowing how to interpret that number is still very difficult. The advantage of FlowWorks is that it gives us a lot of options for various filters we can use to understand the number that’s in front of us.”

 “If you see decreasing concentrations, you might say it looks like the number of cases is on its way down,” he adds. “If you don’t adjust that for the flow rate, then you’re going to come to the wrong conclusion. That’s where FlowWorks really shines – it allows us to take what happens in the collection system into account when we’re interpreting these numbers.”

Aaron Bivins, Ph.D., P.E., University of Notre Dame

 For instance, the concentration might go down for two days, but he can see that the wastewater flow was up due to rainfall, Bivins says. “If you see decreasing concentrations, you might say it looks like the number of cases is on its way down,” he adds. “If you don’t adjust that for the flow rate, then you’re going to come to the wrong conclusion. That’s where FlowWorks really shines – it allows us to take what happens in the collection system into account when we’re interpreting these numbers.”

WBE helps strengthen the relationship between the water sector, clinical health practitioners and policy and decision makers to build a platform of information knowledge-sharing that could proactively assess other potential human health threats, as noted in a Water Research Foundation paper. The foundation recently hosted a virtual International Water Research Summit on Environmental Surveillance of COVID-19 Indicators in Sewersheds examining best practices, standardized procedures and data use. The foundation also has funded a study to do a methods comparison and the U.S. Centers for Disease Control and the Environmental Protection Agency are also doing work on methods development, notes Bivins.

Using WBE findings, a municipality may be able to identify a neighborhood with a high rate of RNA showing up in a particular catchment area such as schools, nursing homes and prisons. “It may be because of the socioeconomic makeup where people are disproportionately affected, the types of jobs people have, their access to health care or the number of multi-generational families who live in that part of town,” says Marsh. “That gives policymakers and public health people the opportunity to drill down and understand where the problem is and maybe that’s where the vaccines should be given first. Or you would use the information to better inform things like mask wearing. You could communicate that this part of town has got a bigger problem with the pandemic right now than the rest of town and needs to be more vigilant.”

Even after a vaccine comes to market, “it makes it even more important for people have systems like this in place because you’re going to want to know if there is SARS-CoV-2 showing up in the collection system in your community after vaccinations take place,” says Marsh. “If the numbers don’t trend downward, that might be an indicator that the vaccine is not particularly effective or not enough members of the community are being immunized.”

Presently, the cost of WBE sampling for COVID-19 is artificially high because it’s not been brought to scale yet, Marsh points out. As techniques are refined and costs go down, Marsh sees the potential for every mid-sized and major municipality in the country to measure the presence of the virus in their communities. “I am confident that if this SARS-CoV-2 Wastewater Based Epidemiology platform can be scaled, it will save lives and provide the data necessary to inform the pace at which commercial activities should be eased or restricted,” notes Marsh.

Beyond the current pandemic, Marsh envisions a time when wastewater analysis will be done at scale as an established operational procedure for virus testing as well as identifying other emerging community health threats.

Kevin Marsh, FlowWorks, Inc. President

Beyond the current pandemic, Marsh envisions a time when wastewater analysis will be done at scale as an established operational procedure for virus testing as well as identifying other emerging community health threats. That will take federal funding and “public health is notoriously underfunded. Municipal utilities also are notoriously underfunded,” notes Bivins. “If we’re learning anything from this pandemic, it should be a few hundred million dollars of prevention might be worth a trillion in avoiding harm and suffering to our economy.”

In order to achieve optimal results, the data has to be reliable and meaningful – the focus of current research, says Marsh.

“Once you’ve got that, then you still have to convince politicians that this reliable and meaningful data should be part of their public health policy,” he adds.  “You need buy-in from the institutions you’re working with and in some cases, that could be a municipality. Those challenges also have to do with crossing the silos that exist within a municipality. For instance, the wastewater people running the plant sometimes are not the same people in charge of the sewers.” Those workers typically do not communicate regularly with public health officials, Marsh notes. “In general epidemiologists are not in regular communications with the collection systems – that’s never been something they needed to worry about,” he adds. “On top of that, some elected officials are reluctant to acknowledge the data “because they don’t want their constituents to think anything bad is happening,” Marsh says.

Even so, Bivins sees great potential in using software such as what FlowWorks offers for WBE. “I think we’ve made a lot of progress,” he adds. “It will be exciting to see where the field goes. I think this is just the tip of the iceberg as far as what’s possible.”

Carol Brzozowski is a freelance journalist based in Pompano Beach, Florida.
She frequently writes about environmental, educational and healthcare topics.