I envision a world where anyone can directly act against new biological threats through surveillance and testing.

It is time to try to build that world. Specifically, I'm building an open-source detection protocol that uses automation and nanopore sequencing.

While this protocol can't be deployed by anyone, I believe it is a step in the right direction towards cheaper and more scalable diagnostic test methods.

Nanopore detection of COVID-19

I have adapted Octant's SwabSeq protocol and Anne Wyllie's Saliva detection to operate on a Nanopore DNA sequencer. I call this method NanoSavSeq (Nanopore Saliva Sequencing).

NanoSavSeq is:

  • Simple (One robotic station)
  • Inexpensive to deploy ($25,000 in capital costs)
  • Cheap to operate (less than $5 per sample)

This method is fast (~6hrs) and can be deployed with a single station, scaling up and down with the needs of the user.


Direct detection of viral particles is usually performed with quantitative PCR. However, this method is difficult to scale, which has caused scientists to look into using sequencing technologies for diagnostic testing for COVID-19.

However, these protocols use expensive sequencers that are difficult to deploy into local areas for testing. I believe that Nanopore sequencers are the solution to the deployment problems of sequencers, while being a fundamentally better technology than qPCR for surveillance and infection tracking (because so many things can be run in parallel and genome sequence can be derived from sample).

When it comes to testing, there are a ton of OTHER things to think about other than the biological protocol: how to get samples, how to manage data, how to interact with regulators, etc. This document is specifically for the technical aspect of deploying Nanopore sequencing tests.

This method should be simple enough for a lab to get up and running within 1-2 weeks, with the ability to scale to as many robots as they can fit. If there is one metric I aim to optimize on, it is simplicity in implementation and use.

I have choosen to support NanoSavSeq on Opentrons robots. By using a robot, I hope to limit the amount of tacit knowledge necessary to run the protocol. By lowering capital costs and giving automation first-class support, I hope companies that need to scale will try NanoSavSeq.

Prior work

The Artic Network has fantastic protocols on sequencing with COVID-19, and the Wuhan Dgensee Clinical Laboratory Co has published a preprint on using using Nanopore for COVID-19 detection. I take inspiration from this work, but am not directly reproducing either one.

Octant has made SwabSeq, and Howard Salis's group has made a Massively parallel diagnostic test. I take inspiration from SwabSeq's method of sample collection, but am adapting their initial protocol to use saliva detection. I am worried about the complexity of primer contamination in Howard Salis's protocol, so am fine with using a PCR machine on an Opentrons deck as a limiting factor to make things simple.


There are 3 major lab steps:

  1. Sample Collection
  2. Automated preparation
  3. Sequencing

There is 1 major analysis step:

  1. Sequence Analysis