Since July, we’ve made major strides in scaling CASPER, expanding Zephyr, and automating our analysis systems. As always, if you have questions or see opportunities to collaborate, please let us know; we’re eager to work with others thinking along similar lines.
Wastewater Sequencing
Our collaboration with Marc Johnson, Dave O’Connor, Rachel Poretsky, Jason Rothman, and others now has a name: the Coalition for Agnostic Sequencing of Pathogens from Environmental Reservoirs (CASPER). The collaboration has expanded to 31 sampling sites across the US, in 19 different cities, and sequences about 60B read pairs weekly. With this much data coming in, we’re increasingly able to detect existing pathogens early enough to be useful to public health, including a detection of measles in wastewater from Kauaʻi County, HI and a detection of West Nile Virus in Missouri. We’re preparing another round of SRA uploads, under the same PRJNA1247874 accession as earlier CASPER uploads.
PHC Global has renewed our ANTI-DOTE subcontract contract for another year, expanding to five US military facilities in the Indo-Pacific region. With this renewal, we’re no longer sequencing wastewater with Oxford Nanopore (ONT) due to the higher cost per base pair.
We ran several experiments comparing different approaches to ONT library preparation for wastewater samples, and found that a modified version of ONT’s Rapid SMART protocol performed substantially better than their Native Barcoding protocol. Even with this higher performance, however, we haven’t been able to make it cost-competitive with Illumina, and have deprioritized further work on ONT for wastewater.
Pooled Individual Sequencing (Zephyr)
We’ve continued scaling up our Boston-based swab sampling program, which we’re calling Zephyr. We now have four field samplers, out 4-5 days a week, collecting around 100-200 samples per day. We plan to maintain this pace throughout the winter and are excited to build out a detailed picture of pathogen prevalence in a major city.
We’ve made two changes to our sample processing approach since the last update:
Our original approach for collecting swabs was “dry”, where the swabs would sit in a container until we began processing them in our lab, potentially as much as five hours later. To improve pathogen recovery we’re now moving swabs into viral transport medium in the field, aiming to keep their dry time under 30 minutes.
As we scaled up, we had been processing increasingly large pools, but we’ve now switched to pools of 40-60 swabs. While we still believe larger pools maximize detection sensitivity per dollar, at our more exploratory stage, smaller pools are valuable because they give more information about pathogen prevalence and system performance.
Over the next few months, we plan to begin experimenting with bacterial sequencing in addition to our current viral-based approach.
The progress of Zephyr can be tracked on our dashboard. The dashboard includes sequencing reads from human-infecting viruses ready for download.
Analysis of Sequencing Data
As we’ve scaled up our sequencing, our metadata system wasn’t up to the task. We’ve redesigned it to be more scalable and easier to work with, and have gone through the majority of past deliveries to get them into a consistent format.
The first focus of our alerting system was engineered viruses, where we saw the largest gap. We recently expanded it to alert us when our viral identification pipeline detects known non-endemic viruses. We’ve also put extensive effort into making the pipeline faster, more robust, and generally more capable. This includes least-common-ancestor analysis, BLAST validation of representative sequences from each detected viral species, and many performance improvements.
Our Outward Assembly is now helping us dig into flagged junctions, and we shared our experience with one production investigation.
We’ve begun a project to automate flag analysis, applying frontier LLMs to replace first-round human review. This will allow our human reviewers to focus on the cases that need human review, and on extending our automated analysis capability. We also hope that this will allow us to increase the sensitivity of the earlier stages of our flagging pipelines, as we’ll be much less limited by human review capacity.
The sewershed CASPER has been monitoring the longest is Columbia MO, collecting over 18 months of deep sequencing data. The first thorough analysis of this data is now available on medRxiv: Untargeted longitudinal ultra deep metagenomic sequencing of wastewater provides a comprehensive readout of expected and unexpected viral pathogens.
Our paper, Inferring the sensitivity of wastewater metagenomic sequencing for early detection of viruses: a statistical modelling study, has now been published in The Lancet Microbe.
Strategy and Policy Developments
As we build out our analysis, automation, and alerting capabilities, we’re encouraged to see metagenomic sequencing gaining traction as a pathogen-agnostic biosurveillance tool to support both public health and biosecurity. The President’s FY 2026 Budget proposes a $52M allocation to CDC for Biothreat Radar, a new pathogen detection system. In a recent blog post, we modeled how such a system could integrate metagenomic sequencing to detect both known and novel pathogens, illustrating one potential path towards nationwide biosurveillance.
We and several other CASPER collaborators are working with Nikki Romanik at Brown University’s School of Public Health as she develops BioRadar, a city-by-city pilot of an integrated biosurveillance system that combines environmental, clinical, and behavioral data to detect emerging health threats.
Organizational Updates
This fall we expanded both leadership and technical capacity, adding expertise across laboratory science, partnerships, and detection response:
Kelly Chafin, Head of Biosecurity Response. Kelly joins SecureBio following a distinguished career in the US government, including senior roles within the Executive Office of the President, National Security Council, Office of the Director of National Intelligence, Defense Intelligence Agency, and Department of Defense. She will lead SecureBio’s efforts to establish and formalize escalation frameworks for detection events and build strategic partnerships with key biodefense stakeholders.
Chris Doering, Genomic Biosecurity Scientist. Chris comes to us from Michael Laub’s lab at MIT where he studied the anti-viral immune systems of bacteria and the viruses that infect them. Leveraging his time spent studying uncharacterized bacterial viruses, he will be characterizing flagged sequences to monitor for emerging and engineered viral threats, and helping us improve our automated classifiers. Outside of work, he loves skiing and rock climbing.
Siham Elhamoumi, Head of Partnerships. Siham joins Securebio with extensive experience deploying and scaling biosecurity technologies and programs globally. She will lead growth and strategic partnerships across biodefense and public health. Previously, Siham served as Director, Strategy and International Partnerships at Ginkgo Bioworks, and was the Program Manager for the Sentinel program at the Broad Institute. She remains an affiliate of the Viral Genomics Group at the Broad where she researches data-sharing and benefit-sharing frameworks in pathogen genomics.
Michael Gomez, Laboratory Technician. Michael joins us from the Sabeti Lab at the Broad Institute of MIT and Harvard, where he helped develop and optimize assays for community biosurveillance and point-of-care testing in low-resource settings. He also contributed to building and scaling a statewide program to routinely screen all Massachusetts dairy farms for H5N1 avian influenza. At SecureBio, Michael will be part of our wet lab team, helping to execute and refine our untargeted metagenomic testing of wastewater and nasal swabs.
James Kremer, Head of Laboratory Science. James joins us from Terrana Biosciences, where he was part of the leadership team driving the discovery and engineering of novel self-replicating RNAs for sustainable agriculture. He has previously led pipelines at Joyn Bio (a Bayer–Ginkgo joint venture) and programs at AgBiome, applying deep metagenomic sequencing, isolation, and automated assays to uncover new biology and develop new technology from environmental microbiomes. James earned his PhD from Michigan State University under Prof. Jim Tiedje and Prof. Sheng Yang He. In his spare time, James loves playing guitar, woodworking, and exploring nature with his 1-year-old daughter.
We recently said goodbye to Research Analyst Harmon Bhasin, who left to co-found a stealth startup, and we’re now hiring for a Bioinformatics Engineer in Cambridge, MA.