Unveiling SwabSeq. We repurposed a part of @OctantBio's platform to detect #SARSCoV2. We think the method is cheap (~1$), sensitive (1-6 molecules), quantitative (3-4 logs), and scalable (10k/day w/o automation; much more with). Lots of details below, I’ll lead with data [1/14]
I’ll take you through a summary of the method and it’s promise and potential pitfalls that lie ahead. To read more, we have provided an overview, detailed protocols, datasets, data summaries, reproducible code, and much more on the Notion page. [2/14] https://www.notion.so/Octant-SwabSeq-Testing-9eb80e793d7e46348038aa80a5a901fd
As a part of our platform, we optimized in-lysate RNA amplicon sequencing to be sensitive, quantitative, low touch, and cost-effective. Two weeks ago @emjbio and @warpcorebreach set out to see if we could repurpose this part of the platform for #SARSCoV2 detection. [3/14]
SwabSeq is simple. Put swabs directly in lysis buffer, then do a one-step RT-PCR for ~40 cycles, pool and sequence. Crucially, we spike in a synthetic RNA that has the same overlaps, but differs by a few bases to tell the difference by sequencing. https://www.notion.so/Synthetic-Spike-In-63f96128c45b4440b11554fae09de950 [4/14]
This spike ensures negative samples still amplify and to go to endpoint, so no secondary normalization is required. In addition, the ratiometric aspect of the test makes it more quantitative and less susceptible to perturbations caused by dirtier samples. [5/14]
We tested a few different ways to put swabs not in the normal viral transport media, but in TE first and then lysis buffer before testing, or directly in lysis buffer, and both kept samples stable for days. [6/14] https://www.notion.so/Swabs-directly-into-TE-or-lysate-58d9000c43f74b2490b7c220b316c193
We had many hiccups along the way. Eg, we discovered one of our primers used to make in vitro templates for N1 was likely contaminating ALL of the index primers made on the same day, and caused a big problem. [7/14] https://www.notion.so/Primer-Contamination-Issues-dfde6628ff8d41f1a49d3141036b0ad1
After about a week of debugging a few issues, it became clear the method was performing very well with a low limit of detection, and quantitative over a large range of viral genome concentration [8/14] https://www.notion.so/Limit-of-Detection-08b9c962874b4772ba27cb1f4923b699
@nathanlubock, @ScottMadScience , @JoshSBloom
helped get our code shareable in the last 24 hrs, as well as docker images, exploratory analyses, etc. For example, this page shows how to get the figure at the top of this thread and much more [9/14] https://github.com/octantbio/SwabSeq/blob/master/analyses/example/example-analysis.md
helped get our code shareable in the last 24 hrs, as well as docker images, exploratory analyses, etc. For example, this page shows how to get the figure at the top of this thread and much more [9/14] https://github.com/octantbio/SwabSeq/blob/master/analyses/example/example-analysis.md
Lots more in the Notion including protocols, code, notebooks, backup designs, some cool features & foolish mistakes, & thoughts on improvement.
to @notion_hq for providing this resource for free & @SlackHQ for bringing together folks thinking about Covid19 Testing [10/14].

There are certainly more challenges ahead. I worry some about how heterogeneous swabs are, and things like that, but nothing deal-breaking. I worry most though about the logistics of swabbing millions of people and reliably returning results... [11/14]
We know we can't do this alone. There are a many harder problems than the tech, from logistics to regulation, but we've face harder challenges before. We are releasing this work under the #OpenCOVIDPledge & happy to enable anyone to get the method working themselves [12/14]
Really proud of this @OctantBio team for pulling together to help with this (& @MollyGasp who helped galvanize us to even try), but also their efforts over the last several years that made this work possible to do in a two week period. [13/14]
Finally, wanted to thank those that took time to help guide us: @lea_starita , @JShendure , @UrnovFyodor , @zhangf, @jonathan_flint1, @leonidkruglyak, et al. And to the healthcare workers putting their lives on the line every day, from the bottom of our hearts, thank you. [14]