Wednesday, May 13, 2015

Oxford Nanopore's London Calling: Pre-meeting speculation

Oxford Nanopore's London Calling confab starts up in a matter of hours.  Alas, several issues scotched my plans to attend (not only does it promise to be an exciting conference, but I simply love exploring London on foot).  It is worth emphasizing that the MinION devices and consumbables have been out in the wild for not quite 11 months at this time.  In that time, Oxford has dealt with a wide variety of technical and logistical headaches. While performance is still variable, many MAP participants are forging forward and the available tools for nanopore data continue to grow.  London Calling will likely bring a burst of new announcements; Oxford's Clive Brown has been giving talks recently but has promised that exciting stuff has been reserved for the confab.  Below is a set of semi-informed speculations calling out likely happenings, mostly based on Clive's recent presentations and tweets.

Fast Runs.  

Oxford has always talked about being able to modulate the translocation speed of DNA through their pores.  Faster translocation means more throughput, but an easy assumption is that data quality would suffer at least proportionally.  Now Oxford is suggested that this isn't so, and that the pores and associated electronics can support translocation speeds of 500 bases per second (versus 30bps with the current protocols) with similar quality.  More details on this are eagerly awaited, particularly a timeline for when MAP participants can start trying the new mode.

Mark I, Mark II

All participants in the conference will be getting a new MinION device, dubbed Mark I (another reason to rue not getting to the conference!). Details as to the differences between the original devices and Mark I haven't been elaborated, but perhaps will be now.  Oxford has also started talking about MinION Mark II, which will have higher throughput (roughly 6X, according to a presentation given recently), presumably by mounting a proportionally greater number of pores.  

PromethION access program details

PromethION is Oxford's follow-on instrument to MinION, with 96 MinION innards running in parallel on a benchtop device.  Further details on the device and its status are likely, as well as how Oxford plans to run the early access to this instrument and perhaps further clarity on pricing.

Dealing with the raw data deluge.  

With the various approaches to increasing throughput -- faster pores, more pores and lots of flowcells in parallel, data storage and processing are anticipated to become headaches. For many applications, throwing out raw data that yields only short and low quality 1D reads (only one strand read) would seem an obvious start.

Local Metrichor?  

With the current MinION, processing data through the cloud-based Metrichor system is only practical with a fast internet connection; performance boosts or the ganging together in PromethION could cause a problem even at well-equipped facilities.  For the goal of moving MinION into the field, expecting very high bandwidth connections just isn't realistic. Oxford has hinted at delivering local installs of the basecaller; perhaps a timeline will be announced.

Conventional applications.  

Despite the short tenure of the MinION, it has already generated at least fifteen peer-reviewed publications, most of which have found it useful.  Between full publications and presentations at AGBT, a checklist of conventional applications have been demonstrated -- microbial genome assembly (with or without correction via short reads), cDNA sequencing, amplicon sequencing, etc.  Expect to see this trend to continue to grow.  Nanopore reads have issues relative to some of the dominant platforms, but low cost and long reads are catching on.

MinION-specific applications.  

Oxford has long stressed that while their devices should be able to grab a slice of the existing sequencing application market, their real target are novel applications which leverage the portability and real-time sequencing capabilities as well as the ability to halt the device, wash it and reuse it.  Already there are two real time analysis packages launched by MAP users (minoTour and npReader). Oxford has announced that they will be sequencing DNA from Hudson and Thames river water samples during the conference.  Also, MinIONs have been seen going to West Africa to analyze Ebola samples in the field.  While the kit required was a bit more than the backpack I have imagined, the ability to rapidly deploy sequencing nearly anywhere in the world (well, as noted above, anywhere with good Internet) is here now.  

More nanopore-optimized bfx tools

In addition to the real-time monitors mentioned above MAP participants have already generated a number of tools either adapted and optimized for nanopore data or brand new tools.  It would be shocking if nothing new was at least floated at the meeting, and perhaps bringing many of the most enthusiastic MinION users into one locale will spur new tool concepts.

Super-simple library preparation.

Preparing libraries for the MinION currently requires a series of molecular biology steps that resemble many other library preparations -- shear, end-repair, adapter ligation -- plus some nanopore-specific steps of putting on some key bits for Oxford's scheme.  Buried in there are also some bead-based cleanup steps.  Having tried the original protocol last summer, I can say that it should be executable by most people who can work a micropipettor, but there are several artful steps around how to rinse and dry those beads (alas, my library went on a bad flowcell -- yes, it was certainly the flowcells fault!).  The recipe I followed is badly out-of-date  -- though the new one is only a bit more streamlined.  Oxford has for a while touted the concept of on-device library preparation, perhaps requiring only a few reagent additions and removals.  If MinIONs are to become ubiquitous with hobbyists and in educational settings, simplicity will be paramount. 

What you won't see

Many commentators, including myself, have complained about Oxford and other sequencing vendors failing to release raw data to support their claims.  Oxford has taken the line in the past -- and I don't expect them to change -- that nobody would believe it anyway.  Instead, their roster of MAP participants has supplied several passionate open source + open data advocates, resulting in a number of MinION datasets being publicly available. I expect this pattern to continue.

Stay Tuned!

I'll be glued to Twitter the next couple of days, since a number of prolific tweet generators will be in attendance, such as Nick Loman, Ewan Birney, Jared Simpson, Matt Loose and Mick Watson.  Not sure if they've agreed on a hashtag yet.  Several of these individuals also maintain blogs.  I'll try to give a roundup of the activity each day.


Anonymous said...

London Calling @NanoporeConf

Dale Yuzuki said...

Yes Keith a daily round-up would be great for those of use who can't stay glued to the Twitter #NanoporeConf hashtag (although I'll be checking in every so often).

In particular I'd be interested to see library-prep improvements to get this crucial step fully field-mobile; much more development is needed before we can go from crude sample lysates to ready-to-run libraries without standard lab equipment.