Methodology and Protocol
Biobot’s Covid-19 wastewater testing product
Lab protocol
Our methods for detecting SARS-CoV-2 in sewage are adapted from CDC protocols. Our approach relies on detecting genetic fragments of the virus that are excreted in stool by qPCR analysis, which does not determine if the virus is dead or active.
Limit of detection (LOD)
The LOD for our lab protocol is 11,400 copies/L of sewage (see more details in Release Notes below). In terms of case estimates, we reliably detect the virus (>99%) when there is at least 1 infected person in a population of 6,500 people.
Data use
Biobot’s wastewater data provides an alternative metric to guide response to the Covid-19 outbreak. We recommend sharing this information with local public health officials. We believe this work has the greatest impact on a statewide level, and hope that you will reach out to your local officials and encourage the expansion of our partnership across your state.
For questions specific to your report, email support@biobot.io
Biobot’s QA/QC protocol
Biobot has an in-house lab facility with a team of scientists dedicated 100% to Covid-19 wastewater testing. All reported data passes our QA/QC protocol:
Sample collection
- 3 x 50 ml samples are shipped with a frozen pack to keep 4C temperature control.
- Documentation collected via online form: location, date, time, flow rate on sampling day, sampling type, precipitation events.
Storage
- Raw sewage samples are received at Biobot and immediately pasteurized. Pasteurized samples are stored at 4C for up to 3 days before viral concentration.
- Extracted RNA is stored at 4C for no longer than 24h before analysis by RT-qPCR.
- Extracted RNA is stored at -80C for the next 12 months.
Sample processing
- 9.6 mL of sewage sample is used for viral concentration and RNA extraction.
- Second and third replicates are kept at 4C for 30 days as back-up.
- Pepper Mild Mottle Virus, PMMoV, is a fecal indicator used as internal control.
- CDC Primers N1 and N2 are used to target SARS-CoV-2.
- Each test primer (N1 and N2) is run in triplicate in the qPCR assay.
- Four positive controls (synthetic SARS-CoV-2 N gene) are run in each 96-well plate.
- Two negative controls (no template) are run in each 96-well plate.
- Standard curves (synthetic SARS-CoV-2 N gene) are run once a week.
Biobot’s data interpretation
Raw viral concentration (genome copies per L of sewage)
The raw SARS-CoV-2 viral concentration is directly measured by the laboratory qPCR assay.
Effective viral concentration (genome copies per L of sewage)
We normalize the SARS-CoV-2 viral concentration to a fecal indicator, to account for differences in dilution. We use PMMoV as this fecal indicator, which is an RNA virus that is commonly excreted in stool. We then derive an adjustment factor that accounts for all other factors we can account for, such as differences in measured concentration due to updates in lab protocols.
Learn more about how the effective concentration is derived here.
Biobot’s Covid-19 incidence estimate
Our latest COVID-19 incidence estimation model is built from Biobot’s dataset. We mined this dataset to derive an empirical relationship between the concentration of virus in wastewater samples and the number of new cases reported in the associated communities. This means that our model provides an estimate of the number of daily new reported cases in your community on the sampling date.
More specifically, we correlate the effective virus concentrations in wastewater samples with the 7-day rolling average of daily new reported COVID-19 cases per 100,000 population in the corresponding communities on the date of sample collection.
In previous versions of our case estimate derivations, our calculations relied on customer-reported flow rates. Our current methodology no longer depends on flow. Instead, we derive our case estimates from effective concentrations, which account for factors including population size, wastewater flow rates, and dilution due to inflow and infiltration.
This incidence estimate represents the projected number of new reported cases that will be reported in your community on the sampling date. This estimate reflects active R&D.
To estimate Covid-19 incidence, we divide the effective viral concentration by a fixed parameter that relates the effective concentration to Covid-19 incidence, as derived from our wastewater dataset:
Biobot’s Covid-19 incidence estimate (%)
To estimate Covid-19 incidence, we divide the effective viral concentration by a fixed parameter that relates the effective concentration to Covid-19 incidence, as derived from our wastewater dataset:
Estimated Covid-19 incidence =
effective SARS-CoV-2 concentration
fixed parameter
Biobot’s Covid-19 case estimate
To estimate Covid-19 incidence, we divide the effective viral concentration by a fixed parameter that relates the effective concentration to Covid-19 incidence, as derived from our wastewater dataset:
Estimated number of new cases in your catchment on the day of sampling =
estimated Covid-19 incidence x catchment population
Our model is built in part on data from clinically confirmed cases. Reported cases include only patients who sought out Covid-19 testing and received a positive test result. This population does not include most people with asymptomatic infections or people without access to testing. Because reported cases are an undercount of the true number of infected people, this estimate is likely an underestimate of the true number of newly infected people in your population. We are working actively to update our model based on the latest science and datasets available.
Release notes
Lab protocol versions
We are continuously working to improve our protocols to increase the sensitivity of our measurements and reduce variability. You can find the protocol that was used to generate your data at the bottom of each page of your report:
| LAB PROTOCOL VERSION | DATE | LIMIT OF DETECTION (LOD) | DESCRIPTION |
| v5.0 (current) | 6/23/2023 | 11,400 copies/L | Bead-based virus capture and extraction with multiplex one-step RT-qPCR for all targets and controls. Improved experimental conditions for increased efficiency, data consistency, and reliability. Preprocessing conditions adjusted per updated CDC guidelines. |
| v4.2 | 10/7/2022 | 9,000 copies/L | Bead-based virus capture and extraction with multiplex one-step RT-qPCR for all targets, including PMMoV and controls. Preprocessing conditions adjusted per updated CDC guidelines. |
| v4.1 | 5/2/2022 | 9,000 copies/L | Bead-based virus capture and extraction with multiplex one-step RT-qPCR for all targets, including PMMoV and controls. Virus capture conditions were adjusted for increased throughput. |
| v4.0 | 02/14/2022 | 9,000 copies/L | Bead-based virus capture and extraction with multiplex one-step RT-qPCR for all targets, including PMMoV and controls. |
| v3.1 | 11/25/2021 | 4,800 copies/L | Kit-based virus concentration and extraction with multiplex one-step RT-qPCR for all targets, including PMMoV and controls. |
| v3.0-3.0.1 | 10/25/2021 | 4,800 copies/L | Kit-based virus concentration and extraction with multiplex one-step RT-qPCR and algorithmic Ct call. Additional controls introduced. |
| v2.3.1 | 09/25/2021 | 3,600 copies/L | Kit-based virus concentration and RNA extraction in duplicate, followed by RNA pooling and one-step RT-qPCR at Biobot and algorithmic Ct calling. |
| v2.3 | 08/25/2021 | 3,600 copies/L | Kit-based virus concentration and RNA extraction with one-step RT-qPCR at Biobot and an improved algorithmic Ct calling method. |
| v2.2 | 07/25/2021 | 2,100 copies/L | Kit-based virus concentration and RNA extraction with one-step RT-qPCR at Biobot laboratory. |
| v2.1 | 06/25/2021 | 1,700 copies/L | Kit-based virus concentration and RNA extraction with one-step RT-qPCR at our MIT partner laboratory. |
| v2.0 | 05/25/2021 | 34,000 copies/L | Kit-based virus concentration and RNA extraction with two-step RT-qPCR at our MIT partner laboratory. |
| v1.0 | 04/25/2021 | 6,400 copies/L | PEG virus concentration and Trizol RNA extraction with two-step RT-qPCR at our MIT partner laboratory. |
Data analysis & model versions
We are constantly iterating on and improving our data processing, analysis, and Covid-19 models to improve the interpretability of our data. You can see which version of our analysis and model was used in this report at the bottom of each page, and you can find more specific details in the release notes below.
| DATA ANALYSIS & MODEL VERSION | DATE | DESCRIPTION |
| v3.1 (current) | 07/25/2022 | We launched an update on the scatterplot, rugplot & hex map to enable better data interpretation and smaller file sizes. Due to an increased volume of samples we are now taking a representative number of 1000 samples over 2 weeks. The representative samples will remain the same across reports weekly. Additionally, minor styling adjustments were made. |
| v3.0 | 02/14/2022 | We report raw and effective virus concentrations. The effective concentration represents the SARS-CoV-2 concentration adjusted for all the factors that we can currently account for. At this time, the effective concentration is the PMMoV-normalized concentration (as in v2.1), adjusted for changes in lab protocol. All plots are generated using the effective concentration. We also updated the case estimation model to use the effective SARS-CoV-2 concentration and a fixed parameter that relates the effective concentration to Covid-19 incidence. Customer-supplied flow rates are no longer required. |
| v2.1 | 11/25/2021 | We report raw and normalized virus concentrations. To derive the normalized concentration, we multiply the raw lab concentration by a scaling factor (reference PMMV divided by your sample PMMV). The reference PMMoV is derived empirically from our entire database. All plots are enerated using the normalized concentration. |
| v2.0 | 10/25/2021 | We launched an updated case estimation model which uses a viral shedding parameter derived from mining our proprietary wastewater dataset. The basic equation is the same as before, but the “viral shedding” parameter is now empirically derived as the amount of virus shed per reported case (rather than based on clinical studies of shedding in stool). As for previous versions, the case estimate is calculated using the raw SARS-CoV-2 concentration, and all other plots are generated using the normalized concentration, normalized with the same method as prior versions. |
| v1.2 | 09/25/2021 | We updated our normalization process for the virus concentration to retain units of copies/L of sewage. We multiply the raw lab concentration by a scaling factor. The reference PMMV is derived empirically from our entire database. As in previous versions, the case estimate is calculated using the raw SARS-CoV-2 concentration and accounts for dilution by using the flow rate provided. |
| v1.1 | 08/25/2021 | We updated detection thresholds to reduce the chance of false positives. Specifically, we’ve raised our limit of detection to ensure that all measurements can be confidently quantified, and are requiring two positive measurements per sample (out of six) to consider a sample detected. |
| v1.0 | 06/25/2021 | Raw viral concentration and Covid-19 case estimates are reported. The model parameter (virus shed per infected person per day) is determined from direct communications with Professor Kyle Bibby and Dr. Aaron Bivins and based on clinical viral shedding reported in Wolfel et al. Nature (2020). |