Tuesday, May 02, 2017

London Calling 2017: A Preview

Oxford Nanopore's London Calling confab runs Thursday and Friday, with a training workshop on Wednesday.  I'll be there -- who can resist a conference nearly at the Tower of London? -- and will also be testing whether my personal "field of nanopore sequencing suppression" can defeat ONT's best trainers.  Here's some preview of what I'll be particularly looking for, though being surprised will be lots of fun too. Much more fun that reading (the wrong) patents!

Bragging Rights

The emergence of near megabase reads this spring, first from Nick Loman and Josh Quick but followed by others, will certainly put focus on ultra-long reads.  I can imagine several categories for bragging rights: longest overall read, greatest read N50, longest chromosome read from within telomere to within telomere (nobody has reported this yet, but reported read lengths are well in excess of the smallest yeast chromosomes).  Capturing an entire bacterial genome in a single read would be a nice coup as well, and eminently plausible given that the second bacterial genome ever sequenced is (Mycoplasma genitalium) has a chromosome of less than 600 kilobases.  The first plenary talk, by Karen Miga, is titled "Linear assembly of a human Y centromere using MinION nanopore long read sequences and that isn't far ahead of Josh Quick's presentation on ultra-long reads, so very early in the conference we may be keyed in long strides in nanopore sequencing -- but don't expect it to end there! Miten Jain isn't until the afternoon.  There's also both a talk and a poster (the latter from Sage Sciences) on the CATCH method using Cas9 to specifically excise a region of interest for sequencing.

The other measure will be MinION yield; has anyone in the field gotten close to the 24 gigabase yields that Oxford has reported internally.  And will Oxford claim to have upped that number?  Oxford has declared a belief that DNA preparation quality is the driver of run yields and that they would be exploring what contaminants are the worst offenders.  If any progress has been made on this point, it will fall on many eager ears.

Status of the Fleet

PromethION will certainly be the subject of chatter from both the company and some of the early access (PEAP) sites.  Live flowcells were apparently shipped in early April, so the community should get some early information on how these are performing in terms of yield and any glitches.

GridION X5 was launched this spring, so expect some early reports of its behavior and perhaps company hints as to how many have been sold.  Since GridION X5 uses MinION flowcells, there shouldn't be any changes there, but it would be hardly surprising if hiccups occurred regarding the built-in computer and when running 5 flowcells at once. 

Clive Brown's recent webcast suggested MinION is unlikely to see any hardware changes for the foreseeable future.  Perhaps we'll see more information on cranking the pores to a higher speed.  Also, the Flongle (see next paragraph) and perhaps full-size crumpet chips (in which the electronics and liquid-facing components are separable) will be discussed, but that 

I think the world eagerly awaits updates on SmidgION, the cell-phone power small cousin of MinION.  I doubt these will be the freebies given out to attendees, given that the product hasn't launched yet, but that would be a bold way to release the product. Flongle is the adapter to allow the SmidgION chips to be mounted in a MinION, which I see as a huge win for the education and hobbyist markets.  Simply having definite release schedules for each, even if those must be padded a bit given the typical slip in such deadlines, would be significant news.

VolTRAX devices have gone out and kits for blitting colored liquids around, but as far as I can tell no actual library prep kits have been shipped yet to early access sites.  

Zumbador is the proposed in-field integrated DNA extraction and library preparation device. Any progress here?

Oxford would seem to an outside observer to have a full plate and to have proposed or operational devices covering a wide space.  But what would London Calling be without new hardware ideas?

Kits and Bits

A sequencing platform isn't just sequencers; the availability and quality of kits are key as well.  Small details can make big difference here.

For example, Oxford library kits and flowcells are currently shipped with cold packs. Getting rid of this "cold chain" would drop shipping costs (which aren't trivial, particularly if you need just one kit) and greatly enhance the ability to operate in the field.  Clive Brown has talked about getting rid of the cold chain, but it hasn't happened.

Similarly, the second version of the rapid 1D transposase prep was originally billed as not requiring any third party reagents and using some clever click chemistry trick.  Apparently this configuration didn't quite click and the kit requires an externally sourced ligase. Is eliminating this dead, or perhaps we'll hear an update.

Around the equinox ONT launched a barcoding kit for the rapid 1D kit, enabling 12 barcodes.  This is very powerful both for dealing with small experiments that don't need a full flowcell as well as enabling rapid method development by allowing multiple conditions (of perhaps sample prep methodology) to be run on the same flowcell.  A dozen is nice, but particularly with PromethION capacities a lot of experiments could go much higher.  Oxford has a PCR barcoding expansion pack enabling 96 barcodes; extending these to the ligation and transposase kits would be useful.  A catch is that with current pricing, rapid libraries cost about $100 in ONT reagents; a 96-sample library could be a budget buster.

Basecalling continues to be an area of rapid evolution, which is good for many people and I'm sure a complete headache for anyone aiming for the clinic.  The newest basecaller, which incorporates some of the "Scrappie" transducer technology, was released in early April; won't be surprised to see some analyses of how well it performs and what the error profile still looks like.  Of particular interest would be understanding the degree to which error is systematic versus random; in the latter case depth can overcome.  ONT has demonstrated successful consensus calling of homopolymer alleles, but this hasn't been released and felt a bit like proof-of-concept; perhaps plans for release will be sketched out.  1D^2 for reading both strands without a hairpin linking them should have hit users just before the conference, so there may be some early looks but it isn't reasonable to expect much detailed independent analysis.  Further details on hardware acceleration of basecalling would be welcome; our group had one nice run recently that spent several days run-end catching up on the basecalling.

Direct RNA sequencing is represented by at least two talks.  ONT has been offering to let more people into the early release of this kit.  The UCSC group just released a preprint with their variation on this protocol to directly sequence bacterial ribosomal RNA

It's been a while since we heard anything about ONT's plans to leverage Cas9 for sequence capture and sequence counting applications.  Andy Heron from ONT has a talk specifically about enrichment methods, so probably we will get updates.

Lots of explicit bioinformatics talks on the agenda and presumably bioinformatics will be key to most of the talks (plus there's Nick Loman's talk, with no title to pre-judge it by).  Many of these tools are out in the wild and described with BioRxiv pre-prints.


Looking at the agenda of talks and posters, the meeting will cover pretty much the full gamut of nanopore applications -- de novo genomes, field use, real time, direct RNA, clinical use, metagenomics, structural variants, methylation -- you name it, it's here.

Well, I'm jet-lagged and should have done even more of this in advance (that I did most of it in advance is quite unusual).  There will be plenty of live tweeting from the conference (I haven't decided whether to live tweet or try to just write everything up).  In either case, I will be sure to analyze the right presentations!


Rafal Marszalek said...

Thanks Kieth, and please keep those reports from London Calling coming (for those of us who cannot attend)!

Torsten Seemann said...

Just write everything up. Plenty of others will be tweeting. You produce eloquent and accurate digests which are very useful to the community. Please do one for London calling!