London Calling: Clive Brown and team plenary
Fri 24th May 2019
Oxford Nanopore CTO Clive Brown's plenary at this year's London Calling was shared with members of his team: Stu Reid, Lakmal Jayasinghe, Andy Heron and Rosemary Dokos. You can watch the video below, or browse the live tweeting at this twitter moment.
Clive Brown, CTO, Oxford Nanopore Technologies, introduced an update on all aspects of nanopore technology – from the pore itself, to sequencing chemistry and instrument software. After presenting the full product range, Clive handed over to Stu Reid (11:38), VP Development, to discuss software and basecalling.
New features to be integrated into nanopore software in the coming months are many-fold, and include data management features, basecalling of modified bases, demultiplexing capability and a pause function on sequencing if the user needs to intervene on a run. Different basecalling options are provided where experimental priorities differ, with a “fast” algorithm for where time-to-result is critical.
Raw and consensus accuracies were also a key topic, as the implementation of the new “flip-flop” algorithm has delivered substantial jumps in sequencing accuracy for both DNA and direct RNA. Another algorithm, Taiyaki, allows users to train their own basecallers, to further optimise their experimental results. Looking to variant calling as a final measure of accuracy, Stu explained that Medaka, software normally for consensus building from basecalls, can now be used to call single nucleotide variants and indels from nanopore data.
Transitioning to Lakmal Jayasinghe (32:18), Senior Director of Nanopore Research, the discussion moved to new pore R10. Soon to be released, R10 is now delivering comparable raw read accuracies to R9.4.1, but an elevated consensus accuracy, with much more optimisation to come in terms of run conditions and basecalling algorithms to push throughput, accuracy and sensitivity even further.
Sample chemistry also plays a fundamental part of nanopore sequencing, and Andy Heron (44:58), Senior Director of Advanced Research, spoke about the power of incremental improvements that aggregate to substantial gains in sequencing performance. In particular, Andy picked up on some key highlights – from direct RNA sequencing (“a new frontier in genomic applications”), to qualitative and quantitative cDNA sequencing, to a multi-horse race for developing single molecule consensus methods.
Finally, Rosemary Dokos (1:08:38), Senior Director of Product Management, summarised the timelines for delivery of new features, including the roadmap for making bioinformatics with nanopore data more accessible, before Clive summarised his future projects: a device codenamed Plongle and ideas for DNA writing.
|00:00:00||Clive Brown's introduction|
|00:03:47||Releases and upgrades|
|00:11:37||Stuart Reid's talk begins: software, accuracy|
|00:15:35||MinKNOW updates: basecalling options|
|00:17:07||Accuracy: raw read|
|00:28:50||Accuracy: variant calling|
|00:32:16||Lakmal Jayasinghe's talk begins: nanopores and accuracy|
|00:33:24||Consensus sequencing accuracy, R9.X vs R10|
|00:37:51||R10: improvements, Qscore updates|
|00:41:12||Future nanopore improvements & updates|
|00:43:01||The route to Q60|
|00:44:56||Andrew Heron's talk begins: sequencing chemistry|
|01:01:09||Single molecule consensus|
|01:08:35||Rosemary Dokos's talk begins: upgrades, timelines|
|01:09:01||Upgrades: Flongle, Mk1C, GridION, PromethION|
|01:11:03||Price per Gb|
|01:11:30||R10 upgrade path|
|01:13:23||2019 upcoming releases & upgrades|
|01:14:17||"And another thing": Clive introduces Plongle|