We had a productive Fall, and are looking forward to an even more productive 2025. Note that this update is a bit smaller than our last update: we’re now targeting quarterly updates.
We continue to be very enthusiastic about collaborating with others in this space: if you have questions or see opportunities please let us know!
Wastewater Sequencing
We’ve scaled up our collaboration with Marc Johnson’s lab at the University of Missouri (MU). They’re now sequencing wastewater from eight sewersheds across four metropolitan areas, with the addition of Riverside CA (in collaboration with Jason Rothman) in December. We’ve expanded to a NovaSeq 25B run every other week, for a total of 136B read pairs in 2024. Marc is hoping to make most of this data public in the next few months, but in the meantime if this would be useful to you please get in touch and we may be able to share.
We’ve recently begun collaborating with PHC Global on their ANTI-DOTE contract. For the next 36 weeks we will be performing deep metagenomic sequencing on four marine blackwater samples each week, and analyzing them for novel pathogens.
In Fall 2023 we partnered with CDC’s Traveler-based Genomic Surveillance program and Ginkgo Biosecurity to collect and sequence both pooled airplane lavatory waste and municipal wastewater influent and sludge. We’ve submitted a full set of aliquots to MIT’s BioMicroCenter for high-throughput library preparation, and will be sending the libraries to Broad Clinical Labs for sequencing later this quarter.
We’re still working on finalizing the data paper we are preparing in collaboration with Jason Rothman, Katrine Whiteson, and SCCWRP on sequencing Southern California wastewater. In the meantime, we’ve made the data available on SRA under accession PRJNA1198001.
Pooled Individual Sequencing
We iterated on the methods for Nanopore sequencing of nasal swabs, and detected multiple human-infecting viruses in our samples. Based on these preliminary results, we plan to expand our sampling efforts in the coming quarter to gather sufficient data for evaluating whether to scale up or discontinue this sampling approach.
Nucleic Acid Tracers
We ran a small experiment to check whether it might be possible to perform a deposition experiment with our nucleic acid tracers using our regular untargeted metagenomic sequencing. We thought the most likely outcome was that we would learn that the quantity we’d need to deposit to be detectable at reasonable cost would be prohibitive, and this was, unfortunately, what we found.
We spiked two of our regular 35mL municipal wastewater samples with 3e6 copies/mL of tracers before sending them to MU for sequencing. They were each sequenced to a depth of 1.0B read pairs, and we observed the tracer barcodes 43 times in one sample and 117 times in the other, for relative abundances of 4e-8 and 1e-7. We estimate that in a sewershed of 100k people you might need to deposit 1e14 to 1e15 copies to be detectable with a NovaSeq X 25B run. This is several orders of magnitude more than we can produce in our lab, and would also be too expensive to produce commercially.
We continue to think these tracers are promising for characterizing wastewater systems using targeted detection methods, such as qPCR and amplicon sequencing.
Analysis of Sequencing Data
When our junction-based detector flags something for human review, the output is a cluster of reads containing a surprising junction. To give analysts more context on potential threats, this fall we developed a tool that performs seeded assembly on very large data sets, allowing us to assemble outward from a set of flagged reads without needing to perform a full metagenomic assembly.
This fall our Robust Detection team investigated methods for detecting growth in a virus’s abundance in our wastewater metagenomic sequence data. We call this approach reference-based growth detection because it relies on a database of viral reference sequences. This quarter, our Near Term Detection team will put this work into production as a reference-based growth detector, our second operational detection method, while the Robust Detection team explores extending the approach to detect growth without relying on reference sequences.
We performed a spike-in experiment to evaluate the real-world performance of our genetic engineering detection system. Our collaborators at MU spiked samples with three different engineered viral particles, and our detection algorithms achieved between 79% and 89% of optimal performance.
Our collaborators Willie Neiswanger and Oliver Liu at the University of Southern California released a 7B parameter metagenomic foundation model, METAGENE-1, trained on our UCI and MU data (preprint). The model is optimized to detect anomalies in short metagenomic reads, and they’re exploring how it can be applied to pathogen-agnostic detection.
The NAO has finished spinning out of MIT. We began in 2021 as a project of Kevin Esvelt’s Sculpting Evolution group, before expanding in 2022 to become a joint project with the newly founded SecureBio research non-profit. This November we completed our journey with MIT by fully spinning out, and are now a project of SecureBio.
We hosted two biosecurity networking events in collaboration with Ginkgo Biosecurity, in October and November, including lightning talks by the Broad Institute, the Boston Public Health Commission, Ginkgo Biosecurity, and SecureBio. We plan to host similar events in the future; let us know if you’d like to be on the invite list!
We’re now fully set up and operational in our new wet lab space, at Tufts Launchpad Biolabs. We’re very happy with our new home, and appreciate the warm welcome we’ve received from the TLB community.
We raised $17k during our Fall Campaign, and we’re grateful to everyone who contributed. We are, of course, happy to have additional contributions, and you can donate on our website.
We’re hiring a laboratory technician to perform nucleic acid extractions and prepare samples for sequencing. This is an in-person role in Boston MA, in our new TLB space. If you know anyone who might be a good fit for the position, please share!
NAO Updates, January 2025
Link post
We had a productive Fall, and are looking forward to an even more productive 2025. Note that this update is a bit smaller than our last update: we’re now targeting quarterly updates.
We continue to be very enthusiastic about collaborating with others in this space: if you have questions or see opportunities please let us know!
Wastewater Sequencing
We’ve scaled up our collaboration with Marc Johnson’s lab at the University of Missouri (MU). They’re now sequencing wastewater from eight sewersheds across four metropolitan areas, with the addition of Riverside CA (in collaboration with Jason Rothman) in December. We’ve expanded to a NovaSeq 25B run every other week, for a total of 136B read pairs in 2024. Marc is hoping to make most of this data public in the next few months, but in the meantime if this would be useful to you please get in touch and we may be able to share.
We’ve recently begun collaborating with PHC Global on their ANTI-DOTE contract. For the next 36 weeks we will be performing deep metagenomic sequencing on four marine blackwater samples each week, and analyzing them for novel pathogens.
In Fall 2023 we partnered with CDC’s Traveler-based Genomic Surveillance program and Ginkgo Biosecurity to collect and sequence both pooled airplane lavatory waste and municipal wastewater influent and sludge. We’ve submitted a full set of aliquots to MIT’s BioMicroCenter for high-throughput library preparation, and will be sending the libraries to Broad Clinical Labs for sequencing later this quarter.
We’re still working on finalizing the data paper we are preparing in collaboration with Jason Rothman, Katrine Whiteson, and SCCWRP on sequencing Southern California wastewater. In the meantime, we’ve made the data available on SRA under accession PRJNA1198001.
Pooled Individual Sequencing
We iterated on the methods for Nanopore sequencing of nasal swabs, and detected multiple human-infecting viruses in our samples. Based on these preliminary results, we plan to expand our sampling efforts in the coming quarter to gather sufficient data for evaluating whether to scale up or discontinue this sampling approach.
Nucleic Acid Tracers
We ran a small experiment to check whether it might be possible to perform a deposition experiment with our nucleic acid tracers using our regular untargeted metagenomic sequencing. We thought the most likely outcome was that we would learn that the quantity we’d need to deposit to be detectable at reasonable cost would be prohibitive, and this was, unfortunately, what we found.
We spiked two of our regular 35mL municipal wastewater samples with 3e6 copies/mL of tracers before sending them to MU for sequencing. They were each sequenced to a depth of 1.0B read pairs, and we observed the tracer barcodes 43 times in one sample and 117 times in the other, for relative abundances of 4e-8 and 1e-7. We estimate that in a sewershed of 100k people you might need to deposit 1e14 to 1e15 copies to be detectable with a NovaSeq X 25B run. This is several orders of magnitude more than we can produce in our lab, and would also be too expensive to produce commercially.
We continue to think these tracers are promising for characterizing wastewater systems using targeted detection methods, such as qPCR and amplicon sequencing.
Analysis of Sequencing Data
When our junction-based detector flags something for human review, the output is a cluster of reads containing a surprising junction. To give analysts more context on potential threats, this fall we developed a tool that performs seeded assembly on very large data sets, allowing us to assemble outward from a set of flagged reads without needing to perform a full metagenomic assembly.
This fall our Robust Detection team investigated methods for detecting growth in a virus’s abundance in our wastewater metagenomic sequence data. We call this approach reference-based growth detection because it relies on a database of viral reference sequences. This quarter, our Near Term Detection team will put this work into production as a reference-based growth detector, our second operational detection method, while the Robust Detection team explores extending the approach to detect growth without relying on reference sequences.
We presented twice at the 2024 CBD S&T conference. We gave an oral presentation on the NAO’s near-term focused work (slides and transcript) and presented a poster on our indoor air sampling review.
We performed a spike-in experiment to evaluate the real-world performance of our genetic engineering detection system. Our collaborators at MU spiked samples with three different engineered viral particles, and our detection algorithms achieved between 79% and 89% of optimal performance.
Our collaborators Willie Neiswanger and Oliver Liu at the University of Southern California released a 7B parameter metagenomic foundation model, METAGENE-1, trained on our UCI and MU data (preprint). The model is optimized to detect anomalies in short metagenomic reads, and they’re exploring how it can be applied to pathogen-agnostic detection.
We’ve also updated our sampling and metagenomic sequencing simulator to include the results of our influenza relative abundance estimates.
Organizational Updates
The NAO has finished spinning out of MIT. We began in 2021 as a project of Kevin Esvelt’s Sculpting Evolution group, before expanding in 2022 to become a joint project with the newly founded SecureBio research non-profit. This November we completed our journey with MIT by fully spinning out, and are now a project of SecureBio.
We hosted two biosecurity networking events in collaboration with Ginkgo Biosecurity, in October and November, including lightning talks by the Broad Institute, the Boston Public Health Commission, Ginkgo Biosecurity, and SecureBio. We plan to host similar events in the future; let us know if you’d like to be on the invite list!
We’re now fully set up and operational in our new wet lab space, at Tufts Launchpad Biolabs. We’re very happy with our new home, and appreciate the warm welcome we’ve received from the TLB community.
We raised $17k during our Fall Campaign, and we’re grateful to everyone who contributed. We are, of course, happy to have additional contributions, and you can donate on our website.
We’re hiring a laboratory technician to perform nucleic acid extractions and prepare samples for sequencing. This is an in-person role in Boston MA, in our new TLB space. If you know anyone who might be a good fit for the position, please share!