The new SARS-CoV-2 test from @nanopore, LamPORE is going live in the NHS this month - here is the news page on this from nanopore. https://nanoporetech.com/about-us/news/oxford-nanopore-technologies-partners-uk-government-roll-out-lampore-new-generation This tweet thread is to explain the science behind this, but first some housekeeping >>

First bit of housekeeping, as my long time followers know (and apologies for the repetition!) I have a conflict of interest: I am a long established paid consultant to Oxford Nanopore (along with some other companies); Oxford Nanopore do not vet my tweets.

The second, big picture, view point is that in the fight against COVID, we need to increase testing 10 to 100 fold worldwide and make it more accessible; LamPORE is a great technology to do this, but not the only one being trialed.

It is far, far more important to me that we stop the viral transmission (itself a stop gap measure until vaccines or treatment) than which particular technologies are part of that solution.

Back to LamPORE. LamPORE is the combination of two established technologies - Loop mediated amplification (LAMP) and nanopore sequencing.

LAMP has been around for a while (since 2000; I remember seeing a live demo at RIKEN in the 2000s) and often considered a curiosity with really quite complex primer design to target the amplification and a notorious "explosive" amplification behaviour

(Primers are DNA sequence of the experimenter's design, and they "flank" the piece of DNA you want to amplify, both in RT-PCR and LAMP; the general term means to "prime" a reaction and it is concept across DNA molecular biology)

LAMP has two benefits here. The first is all it needs is constant heat (isothermal in the trade) which makes the lab kit to do this far easier. The second is that most amplifications go to completion (benefit of "explosive amplification").

This means if you precisely measure the DNA primer amounts in differnet tubes you get the same molarity of resulting amplified products. This is super useful.

The downside of the explosive amplification, which has the potential for cross-contamination is controlled somewhat by the format but most of all that @nanopore cleverly designed their first layer of barcodes *into* the LAMP primer design. This is fiddly.

(LAMP amplification diagrams completely bamboozle me, and I hear that the barcode insertion makes it even more complex)

An brief explainer on "DNA barcodes". DNA barcodes is shorthand for designing schemes where the DNA you introduce has a segment that changes under your own design plus the standard business end. In this case, there are 96 slightly different primer sets, each with a unique barcode

The unique barcode means after the molecular biology processes you can simply mix the resulting reactions. If you use DNA sequencing as a readout then the barcode assigns each molecule to a unique starting reaction. So in this case you can do 96 reactions in parallel

(this is sometimes called multiplexing, though multiplexing is also used for other quite different techniques; DNA barcoding is more specific and ends up being limited only by primer design and robotics)

That's the LAMP bit; now onto the Nanopore bit. Nanopore uses a tiny protein pore (a nanopore) in a lipid bilayer (like a soap bubble) and when DNA passess through the pore in a controlled manner the amount of water ions that can pass by the DNA is specific to the DNA sequence

You can measure this differential passage of water ions relative to time with a sensitive chip and if the single nanopore is placed over one chip sensor - and some v. smart algorithms to sort out the signal - you can work out the DNA sequence that came through the pore.

This is all super-fiddly and complex, and over a while @nanopore have gotten all of it to work. There are some pros and cons vs other ways of determining DNA sequence but the two important ones here is that (a) devices are small and (b) they are super fast.

How small? The actual pores and chips are mini-mini (smaller than many coins). They have made mobile phone compatible devices. But, in fact, to do everything in a sensible package (in particular the computing) their best device looks like a thick DVD player with door.

And you can obviously put many of the "chip stations" side by side, which is their printer sized box. There is also a big daddy machine (a sort of small photocopier sized thing) which is too big for this.

How fast? Once DNA is introduced to the pore it will start reporting sequence within 15 minutes. The standard set up (called a "flow cell") has around 500 active pores (there is a trick about getting one pore per sensor which means this varies a bit).

So - full circle back to LamPORE. Nanopore designed LAMP primers to the SARS-CoV-2 genome with a 96-way barcode. They actually have 4 sets of primers per sample reaction ('well' in molecular biology speak); 3 on the virus across different genes and one human RNA (beta actin)

Once swabs are RNA extracted, the LAMP reaction happens (from 1 to ... 96) and then at this point one can just mix the resulting reactions, add the nanopore adaptors (which enable the DNA to be sequenced) and put on nanopore - up to 96 samples, <90mins

But one can be more flexible here; at the Nanopore adaptor stage one can use a second level of barcodes - their kits have 8-way, allowing 768 final parallel ("multiplexed") samples.

This was all working in May; then one needs to test test test. Test inside of ONT and then test with other laboratories, such as @adbeggs in Birmingham with real clinical samples and PHE labs in Porton Down. It gives very very concordant results with RT-PCR even at low levels

The fact that the readout is DNA sequence means there are unique stretches of DNA in the resulting LAMP products not present in the primer sequence. These are short, but, if present, absolutely clear cut that the SARS-CoV-2 sequence is in the reaction

The Human primer means LamPORE can distinguish "empty" samples (eg, a poorly taken swab) from real negatives (which will have a positive on the human RNA, negative on the virus).

The up to 768 multiplexing (two level bar codes) means the smaller DVD machine (MinION Mk1c) can do up to 3K samples per day; the printer sized machine (GridION) 15K per day. Both are "pop lab" compatible (ie, van portable)

Two GridIONs is close to a lighthouse lab capacity, and the fact that one can make these things portable means that sample logistics options (ie, getting swabs extracted to machines) can have other modalities - bringing the testing to people.

(lots of complications - indeed, logistics is a key thing of people, swabs and reagents for testing overall - but the portability gives more options)

Hats off to the @nanopore crew who developed all of this (in particular @RosemaryDokos and Dan Turner) and of course the whole @nanopore team for this amazing tech (as @The__Taybor says, we're shipping soap bubbles worldwide to measure single molecules).

I will end with another reminder of my conflict of interest: I am long established paid consultant for Oxford Nanopore. Oxford Nanopore do not vet my tweets.