Thread by @josh__levy

Wastewater contains a treasure trove of information for monitoring public health.

In our new paper, we solve several lab and computational bottlenecks to achieve timely, accurate, and cost-effective SARS-CoV-2 surveillance using wastewater sequencing. nature.com/articles/s4158…

Our approach enabled early detection of emerging virus lineages (by as much as two weeks) and often detected lineages not seen clinically. Using a single sample, we profiled virus lineage dynamics from the scale of an entire city all the way to individual buildings.

This project truly took a village. It was a close @K_G_Andersen lab collaboration with @smruthi_karthik of @KnightLabNews, and so many contributors from @UCSanDiego @scrippsresearch @ScrippsHealth @radychildrens @sharphealthcare @SanDiegoCounty @CAPublicHealth @my_helix

This effort started in the early pandemic. @smruthi_karthik , @KnightLabNews , and many others at @UCSanDiego installed a campus-wide wastewater sampler network, enabling daily measurements of the prevalence of #SARSCoV2 at >100 individual sites. returntolearn.ucsd.edu/dashboard/inde…

Obtaining quality virus sequencing from wastewater is no simple feat, and protocol modifications made by @smruthi_karthik and others @UCSanDiego enabled massive improvements in sequencing quality and opened the door to new computational analyses.

Upon sequencing #SARSCoV2 in wastewater we observed considerable virus diversity, suggesting that sewage samples often contained multiple distinct virus lineages.

To estimate frequency of #SARSCoV2 lineages, we developed Freyja, a tool that uses mutational "barcodes" of all virus lineages in order to "demix" samples. Freyja is robust to incomplete sequencing coverage, even for coverage as low as 50-60%. github.com/andersen-lab/F…

After benchmarking, we tested Freyja on real wastewater sequence data and found that it enabled early detection of #VOCs as well as extended periods during which these variants circulated on campus while managing to evade clinical detection.

We found that we could more generally profile virus lineages beyond these VOCs, and that our wastewater surveillance captured similar lineage frequencies to what we observed clinically.

At the largest #WW plant here in San Diego, we observed a transition from a diverse array of Delta lineage viruses to a single early Omicron lineage (BA.1, detected in #WW 11 days prior to clinical), which was eventually overrun by its more fit descendant BA.1.1.

By combining wastewater viral load measurements with our estimates of relative lineage abundance, we were able to estimate the magnitude of viral load due to each individual lineage.

Perhaps most importantly, we continue to make all of this data available to the public in real-time, and use fast-turnaround sequencing to go from sample collection to results in just over 24 hours. SEARCH Dashboard (H/T @akaEscapePlan): searchcovid.info/dashboards/was…

We’ve made Freyja free to the public via a range of platforms, and with the help of @kevinlibuit and the folks at @theiagen, it's also available via @TerraBioApp, for those who prefer to use a web interface and/or cloud computing resources. Workflows: dockstore.org/search?entryTy…

It’s been awesome to see groups around the US and the world using Freyja for their analyses (like @fishtiaq and @KGBscience). @BiobotAnalytics and @US_FDA have some incredible dashboards that use Freyja under the hood.
biobot.io/data/
fda.gov/food/whole-gen…

We are still developing the Freyja package, which now includes features like bootstrap-based uncertainty estimation and outputs like fully functional wastewater dashboard web pages. And we've got lots of new functionality coming soon!

Freyja is also available via conda
anaconda.org/bioconda/freyja
and docker (via StaPH-B)
hub.docker.com/r/staphb/freyja