Thread by @nanopore

And now at #nanoporeconf, an update from the team @RosemaryDokos, James Clarke and Stu Reid. A thread

First up, James Clarke with an introduction to @nanopore sequencing. More than 1,000 publications use nanopore sequencing nanoporetech.com/resource-centre, and we estimate that ~1.7 petabases of nanopore sequence has been generated to date #nanoporeconf

JC: Here’s how nanopore sequencing works. Library prep straightforward: add an adapter to each end of the sample DNA molecule, containing leader, motor protein, motor stall, tether #nanoporeconf youtube.com/watch?v=RcP85J…

JC: the @nanopore signal is rich, decoded in real time to elucidate sequence. Updating basecallers has resulted in profound performance iteration over recent years #nanoporeconf

JC: a key feature of the ASIC is that it can reverse the potential in real time - this means we can unblock and eject DNA to keep the pores running, and also enables adaptive sampling in real time #nanoporeconf

JC: latest performance gives high accuracy, short-to-2.4Mb continuous reads, direct methylation analysis, native RNA sequencing, high data yields and fast time-to result #nanoporeconf

JC: A range of nanopore sequencing devices provides for sequencing at any scale – whether on-demand smaller tests with Flongle, or high throughput human/plant genomes on PromethION nanoporetech.com/products #nanoporeconf

JC: In the MinION Access Programme, as the yields from a MinION flow cell went up we realised people don't always need or want that much data. That's where Flongle comes in - an on-demand, low cost flow cell. #nanoporeconf

JC: a few months ago we opened a new manufacturing facility, which allows us to scale up production of flow cells – Flongle in steep production ramp-up
#nanoporeconf

JC: PromethION is our device for highest throughput. It's still on-demand for flexibility (up to 48 flow cells can be used at once). PromethION flow cells are different to MinION/GridION: 3,000 channels. PromethION includes powerful GPUs #nanoporeconf nanoporetech.com/products/prome…

JC: PromethION Flow cells had more variability in quality than we would have liked in 2019, but production now scaled on optimised versions – with these we have reached 8Tb on a single P48 run using MinKNOW MK4.0 plus LSK110 (fuel fix)
#nanoporeconf

JC: all the improvements we deliver make an incremental improvement so please do make sure you keep your software and firmware updated! #nanoporeconf

JC: electrode connection & well structure key in PromethION flow cell QC - fix by re-adjusting the lithography process-now very tightly controlled, FCs much more consistent. We hit 8Tb run on P48, we're looking forward to seeing how these FCs perform in your hands #nanoporeconf

Now, Stu Reid: SR: Development of machine-learning basecallers has driven accuracy enhancements. “flip-flop” RNN base caller is now in MinKNOW 4.0 / Guppy 3.6, for single read modal accuracy of > 97 % - you can recall old data #nanoporeconf nanoporetech.com/accuracy

SR: a range of basecallers are in research, providing multiple pathways to high accuracy data with different features #nanoporeconf

SR: Barcodes are designed for edit distance, complexity, GC content. Choice of PCR (96), Native (24) and Dual (2304) Under the most stringent settings you can achieve 99.999 % correct classification accuracy. #nanoporeconf

SR: For high @nanopore consensus accuracy, use the best tools – updated Guppy, Canu/Flye, Medaka is recommended tool. R9.4.1 ~Q45, R10.3 ~Q50
for microbial community #nanoporeconf nanoporetech.com/accuracy

SR: need to be careful about how some people show their single molecule consensus charts - there are no filters on this genome browser shot. Consensus levels of accuracy at the single molecule level >Q20 at 2X depth with Bonito pair decoding. #nanoporeconf

SR: @cyclomics have a sophisticated high accuracy single molecule workflow, important to have high resolution for liquid biopsy application #nanoporeconf

SR: you will hear more about high accuracy, contiguous human genome assemblies with nanopore PromethION at #nanoporeconf - we have seen strong BUSCO scores alongside high accuracy data #nanoporeconf

SR: For methylation analyses, simultaneous, real time analysis of 5meC and 5hmeC is now possible: trained model now available. Per-read annotation and multi-read aggregation with Megalodon
#nanoporeconf

SR: we’ve developed a completely new touch screen UI for MinION Mk1C, and now all devices #nanoporeconf

SR: new features in updated @nanopore UI include ability to run sequence alignment via the UI, combine with barcoding for powerful visualisation, adaptive sampling enabler. Tutorials and management much improved #nanoporeconf

SR: LIMS and data management greatly improved with MinKNOW 4.0. Sample sheet uploads already enabled, adding sample by barcode soon. VBZ compression is now on by default, for 35% smaller files and using <10% CPU time of gzip #nanoporeconf

SR: Adaptive sampling (Read Until) will run in MinKNOW 4.1 in Q3 (GridION, Mk1C), in dev for PromethION
#nanoporeconf - you can hear more tomorrow from @mattloose too
youtube.com/watch?v=fwHreH…

SR: we provide a selection of tools for downstream informatics. EPI2ME simplest: cloud based, no IT infrastructure needed. EPI2ME labs suited to intermediate users and includes tutorials. Opensource Github if you want to get your hands dirty nanoporetech.com/nanopore-seque… #nanoporeconf

SR: We're improving output in multiple ways, including reducing fuel depletion and extending total available run time of flow cell #nanoporeconf

SR: reducing fuel depletion with ligation sequencing kit (LSK110) – includes ‘fuel fix’ that increases output on short read libraries & reduces interactions needed on longer read libraries. Supports production-scale seq #nanoporeconf

SR: extend total available flow cell run time by reducing mediator depletion. Concentrations and ratios of mediator determine run time: deeper well design can extend run #nanoporeconf

SR: sensitivity can be increased substantially by getting the analyte closer to the nanopore. Aiming for 200X sensitivity with new trans-tether, so far 10X (eg DNA: LSK109 – 5 Gbases from 1 ng)
#nanoporeconf

SR: looking further into the future... new, smaller, lower power ASIC opens up newer format possibilities eg Plongle (plate Flongle) and SmidgION (smartphone sequencer) #nanoporeconf
nanoporetech.com/products/plong…

SR: new voltage-based platform in research could dramatically increase density of sensors, theoretically as high as 4Tb of sequence on the size of a MinION #nanoporeconf

Up next @RosemaryDokos, VP Product Management, will talk about what’s happening, when #nanoporeconf

RD: a summary of new kit releases #nanoporeconf

RD: we’ve been doing a lot of work to support your high throughput sequencing. As well as XL kits – available now – we have developed automation protocols on Hamilton liquid handlers, and high throughput prep protocols on FastPrep96 #nanoporeconf

RD: you've asked for more barcodes, and you're getting more barcodes! We’re developing 96 rapid barcodes that can be prepped in 10 mins, but in the meantime 96 native barcodes shipping in June. 96 x 96 dual barcodes protocol is in community #nanoporeconf

RD: VolTRAX V2 features are progressing. New UI launching in summer, with live cartridge feedback. User-programming API to developers in summer too. PCR and fluorescence in progress #nanoporeconf nanoporetech.com/products/voltr…

RD: this chart was designed by Spike and Clive many years ago, and its still relevant. Increasing translocation speed continues to be a driver for yield #nanoporeconf

RD: a reminder of the latest accuracy figures. Key updates recently include R10.3, which includes homopolymer resolution, and is now available for Min/Grid, with early access PromethION users starting to use it #nanoporeconf nanoporetech.com/accuracy

RD: these are key releases that will continue to support accuracy increases #nanoporeconf

RD: we have core values when we launch a new product. One is accessibility, and our pricing is designed to enable access to all. #nanoporeconf

RD: pricing is transparent, and competitive per Gb – but not everyone is driven by cost per Gb and cost per sample is also competitive and enabled by multiplexing. You can find the pricing listed here #nanoporeconf
nanoporetech.com/products/compa…

RD: as you get towards 200Gb per flow cell, that's two human genomes per flow cell, and at $500 per flow cell that's a very competitive offering especially when you get rich data as part of the package #nanoporeconf

RD: There are multiple options available for targeted sequencing – amplification based, which are fast and low cost/high plex. Amplification-free methods eg Cas9 which also retain methylation, and adaptive sampling does the targeting on the flow cell in real time #nanoporeconf

RD: when you invest in nanopore, you are getting future upgrades for free, and there is plenty of headroom for growth #nanoporeconf

RD: On RNA sequencing, improvements with kit 10 mean it’s possible to now get millions of cDNA reads. And nanopore remains the only way to sequence RNA directly #nanoporeconf

RD: so many great stories from the metagenomics community -- assembly, river metagenomics and co-infection detection to name just a few #nanoporeconf

Thanks for tuning in -- this update will be available on demand later today, along with all the other talks from #nanoporeconf so far!