Monday, 23 April 2012

Who (if anyone) will be remembered in the race for the $1000 genome?

A journey back in time 
This week 59 years ago, the race to elucidate the structure of DNA was officially classed as won with the publication of James Watson and Francis Crick's "A Structure for Deoxyribose Nucleic Acid" in the journal Nature.
The discovery of the double helix was a landmark in modern science and paved the way for many of the techniques and abilities we now take for granted.  Since 1953, the names Watson and Crick have become synonymous with DNA and genetics.  In fact, if quizzed on who discovered the double-helix, many would offer only these two oft-repeated names.

Neglected heroes
Fewer would be aware of two manuscripts penned by Maurice Wilkins and Rosalind Franklin, which appeared in the same issue of Nature and contained the results of X-ray diffraction experiments that played a major role in revealing the structure of DNA to Crick and Watson.  At the time of publication, the contributions of Franklin and Wilkins were downplayed but in subsequent years were acknowledged and saw Wilkins share the victory of the Nobel prize in Physiology or Medicine in 1962.  Franklin only missed out because of her untimely death from cancer and the fact that the prize could not be awarded posthumously.

The final elucidation of the structure also benefited from knowledge accumulated from years of pioneering science on both sides of the Atlantic.  It is unlikely that Watson and Crick would have arrived at their final conclusions if it had not been for their emulation of Linus Pauling's model-building or Erwin Chargaff's discovery of the fact that purines and pyrimidines exist within DNA in equal proportions.  Work like that of Avery, Hershey and Chase no doubt helped to steer research in the direction of DNA and away from the proteins that many believed must be the units of heredity.  The list of those whose work contributed to the final discovery is a long one and well documented elsewhere.

To the victors the spoils
There can be no question of Watson and Crick's own contributions however.  All science builds on the work of others to some degree and we would not know of them today if it had not been for their own passion, tenacity and scientific flair.  The point here is that there will always be winners and then there will be contributors.  Only one of the two will achieve true glory and be awarded by synonymy with their achievement.  The rest will be demoted from headlines and similarly from the memories of most laypersons.

And onto 2012...
Today in the field of genetics and genomics we speak often of another race - the race for the $1000 genome.  Over the course of the past six decades the face of scientific research has changed immeasurably.  Technology has progressed at bewildering pace and scientific research has become increasingly associated with commercial industry.  If asked who I believe to be the competitors for the $1000 genome, I would answer Illumina and Life TechnologiesOxford Nanopore have caused a buzz with promises of their technology's capabilities but in the absence of public data or full commercial launch I would elect to keep them off my list for now.  I will gladly be proved wrong.  Regardless, it is inevitable that - just as with Watson and Crick - the winners may be remembered, but the many contributors along the way will most likely find themselves in the footnotes.

The invisible (wo)men
One point I wish to make here is that my list of competitors consists of companies rather than individuals.  Even as someone working in science I increasingly find myself linking scientific progress with the names of corporate groups rather than those of scientific researchers.  Gone are the quaint mental images of eccentric gents in white coats toiling for personal glory in university labs.  In their place is an image of ... well... not much really.  The contributions of individuals appear to be lost in corporate smog leaving us to associate the scientific advances with company logos and occasionally a CEO with a love of the limelight.  The concept of companies jostling for marketplace superiority lacks most of the charm of of the Watson and Crick story.

A leap forward?
The other point is the question of whether this particular race will actually produce a memorable winner.  The story of the race for DNA cumulated in one of the most dramatically consequential discoveries of scientific history.  The $1000 genome is a human concept - simultaneously an incentive for development and a yardstick to mark progress.  Furthermore rather than requiring seminal thought or actual discovery, it is something that can likely occur as part of the natural momentum of current technologies.  The real benefits of a $1000 genome versus a $2000 genome or even current costs are difficult to quantify.  Let us not forget that even a $1000 genome is out of reach to many on the planet.

True advances
The reagent-less genome or the $1 genome on the other hand may mark the point where sequencing really is available to everyone, everywhere.  Developments like these will require true discovery and produce benefits regardless of social class and geographic location.  This would mean genome sequencing without wealth or adequate insurance and provide every human on the planet with the right to benefit from the information locked within their own genetic code.  This would represent a true landmark and produce a story our grandchildren learn about from textbooks just as we learned about the double-helix.

Who knows?  Perhaps the story will even feature a protagonist rather than a corporation.

Wednesday, 29 February 2012

Illumina: Aces in the hole?

The past two weeks have seen Oxford Nanopore reclaim the title of most talked-about sequencing technology and the wake of hyperbole, hope and suspicion left behind by the announcement that their GridION and MinION systems will be available this year has done some damage to the field's biggest players - Illumina and Life.

Not only did the announcement of the Oxford bred technology cause the dropping of many a scientist's jaw but it also led to 3.9 percent and 7.8 percent drops in share prices for Illumina and Life respectively.  The Web is awash with questions of what effect nanopore technology will have on the field as a whole but whilst the world awaits something tangible from Oxford Nanopore, let us not forget that Life and Illumina have each made some major announcements in early 2012.

Old toys improved
Illumina announced significant improvements to their desktop MiSeq system in the form of lower runtimes, 250bp paired-end reads, and a dramatic increase in throughput to 7GB per run.  Many had believed that the system's potential for improvement over release specs was limited but these announcements suggested the opposite.  James Hadfield at CRUK has even questioned if the system will in fact prove scalable to the level of human genome sequencing.

Life meanwhile released their 318 chip and its nominal 1GB output was quickly exceeded with internal data runs exceeding 2GB.  Meanwhile, throughput on the smaller 314 and 316 chips continued to increase at an even greater pace.  Plans to launch 2x200bp paired end reads and 400bp single end reads in 2012 were also announced at an AGBT meeting where Life's inventive marketing was impossible to escape.

Perhaps the biggest announcements however came in the form of...

Sparkly new toys

January saw Life's announcement of the Ion Torrent Proton, a $149,000 benchtop analyzer estimated to ship to early access customers in mid-2012 coupled with the Ion Proton I chip - said to be ideal for sequencing human exomes.  Human genome sequencing was predicted to come to Ion Torrent by 2013 with the Ion Proton II chip promising the holy grail of human genome sequencing for $1000, and in a day.

Illumina returned fire the same day by press releasing their HiSeq 2500 system which similarly promised the ability to sequence a human genome in a day (120Gb in 27 hours) but with a hefty $740,000 price tag or $50,000 as a field upgrade to the HiSeq 2000.  Full commercialization of the new system is expected in the second half of this year.

Consumer opinions
Regardless of the fact that each set of announcements could be argued to be equally important or advantageous, the buzz seems to have surrounded Ion Torrent.  Major acquisitions of Illumina sequencers have continued however there seems to be an increasing attitude of indifference and even caution directed at the Illumina technology despite its proven utility and ability.

One online poll shows only 50 percent of correspondents believing that purchase of Illumina technology is still a good idea with 25 percent classifying such purchases as 'very risky'.

These changing attitudes are no doubt partially due to our love of new toys and a superior marketing campaign by Ion Torrent, however less fickle motives also exist.

Illumina beaters

Currently, the PGM offers greater flexibility in terms of run-size.  The existence of the 314, 316 and 318 chips offers the user a level of control over run-size and cost that cannot be achieved on MiSeq.  The announcement of the Proton I and II chips would suggest that this advantage will translate to the higher-throughput sequencing, potentially threatening future HiSeq purchases.

The announcement of the Proton has shown us all that the chip is not necessarily the machine after all however the increases in throughput achieved on Ion Torrent technology have been impressive and there is a belief that this will continue to bring future improvements and advantages long after Illumina chemistry has hit its ceiling.

Both the PGM ($50k vs $125k for MiSeq) and Proton ($149k vs $740k for HiSeq 2500) are priced well below their Illumina-based counterparts.  These prices are based on sequencer only but the numbers speak for themselves.  There have also been reports of Life being more willing to negotiate on cost, despite lower list prices.

There is a growing belief that Life will continue to be the winner in these important areas and that Illumina represents the technology of yesterday.

Surprises in store?
Is this really the case though?  Or do those who write off Illumina as an old man playing a young man's game risk a shock?  Beyond January's announcements could Illumina have something else up their sleeve?

Whilst travelling home from AGBT, on the bus to the airport I had the pleasure of meeting Dan Koboldt who described speaking to a casually confident Jay Flatley during the meeting.  I can personally say that while the buzz at AGBT may have been around Life (and later Oxford Nanopore), Illumina were far from silent.

Talks and posters made it clear that Illumina is still the big player in the market, being used widely in traditional as well as increasingly clinical applications.  Roche's interest in the technology speaks volumes about its perceived clinical utility.

In and around the elegant Illumina lounge between distribution of drinks and genomes on USB sticks there were whisperings of potential new products.  There was talk of new kits that would add greater flexibility to run-size and cost.  Also, there were suggestions of entirely new sequencing technologies arriving within the calendar year.

Whilst the Oxford Nanopore announcement raised questions about the future of the field and the price of shares, it also raised another: how will Illumina's relationship with the company and its exonuclease sequencing technology  manifest in the market?  Beyond that how will it align with Illumina and Oxford Nanopore's other hardware offerings?

There is no doubt that Illumina's position is being increasingly challenged by the emergence of new, innovative technologies but the status of market leader was not achieved by accident.  The prowess that achieved sequencing dominance could still surprise the challengers with a reappearance.

In this market nothing is certain but I won't be writing off Illumina just yet.

Friday, 17 February 2012

AGBT 2012 - Nanopores or NanoBores?

I am writing this anticipated entry on the fly so please forgive apparent illiteracy.

Perhaps the most eagerly awaited event of AGBT 2012 is Clive Brown's revelations about Oxford Nanopore's Strand Sequencing and the plans to commercialise the technology, independently, this year.

Slightly later than  scheduled, at 11.50 am Clive took to the podium and commenced his 20 minute presentation. 

Below is my attempt at a blog entry.  To be honest I was not expecting the bombardment of new information that was presented and the attempt at a report deteriorated into a series of rushed fragments.  Oxford Nanopore's website has now been updated and I recommend having a look there for further details.  There is a real buzz in Marco right now!

He commenced with an overview of the technology's principles (well explained elsewhere). He says over 1000 different types of nanopore have been tested.  Many different processive enzymes and sample preps too.  One prep takes 15 minutes.

Sensors can do up to 1000bps.

Whole PhiX 5 kB genome as single fragment, approx 4% raw read error on single pass (mostly deletions).  10 and 20 KB fragments achieved too.  No degradation of signal over time.

Initial RNA data tested and data looks as good as DNA - to be revisited later.

Sequencing achieved directly from rabbit blood - no prep sample!

MinION - USB sized disposable sequencer to be released this year.  150 MB per hour. Scalable to 1GB.  Life span 6hrs.  Sample goes in and then plugs into a USB port.

GridION 2k output (2012)  25-125GB per day.  GridION 8k (2013) 125-625GB per day.

Genome in 15 mins by next year (50x) with GridION 8k. 

Both to be released second half of this year.

WAY too much info for one talk.  Amazing. 

Wednesday, 1 February 2012

Oxford Nanopore emerge from their burrow

Spring has come early in Oxford.

For anyone who assumed all the big sequencer news for 2012 (Ion Proton / HiSeq 2500) was already announced, Oxford Nanopore have just proven you wrong by announcing that CTO Clive Brown will present their "strand sequencing" method at AGBT in 2 weeks and will "independently commercialize" the technology in 2012.

It is important to note that Illumina, who are invested in Oxford Nanopore technology have rights to commercialize their "exonuclease sequencing" technology.  To my knowledge, the strand sequencing method uses a polymerase to move an intact strand of DNA through a nanopore, whilst the exonuclease removes dNTPs one at a time, for their passage through the pore.**

Not only will it be interesting to hear what stage the technology is at (and what has been sequenced) but how the relationship between the two companies and the two distinct technologies develop will no doubt generate some headlines also.

Clive recently posted ominously that "The days of short reads are numbered."

Are they?

Hopefully I'll find out in Florida!

**It is worth noting that startup company Genia claim their technology will electrically control the movement of the strand through the Nanopore

Tuesday, 10 January 2012

Spurred into action by AGBT and Ion Torrent

It had been my plan to resist posting any further entries here until January's end when I would have been able to share some preliminary work of relevance to anyone working in the area of mutation detection from NGS data, particularly mutations of diagnostic relevance.  That work is underway and will find its way onto this blog soon, however two pieces of relevant (if disparate) news received within twenty four hours of one another encouraged me to debut early.

Firstly came a much hoped for but unanticipated e-mail from the administrative staff taking care of this year's AGBT conference at Marco Island, FL.  I had hoped to attend my first AGBT in 2011 but was a little slow on the trigger and missed out.  I was determined that this year would be different and so sought managerial approval months in advance.  Getting approval so far ahead meant attendance was a sure thing, right?


As so many others also discovered in the year when a scientific conference began to rival even the greatest of rockbands for speed of ticket sales, unless you had managed to access the AGBT website, populate the necessary form elements and execute the requisite number of mouse clicks within 3 minutes of tickets going on sale you were too late.  A trickle of tweets immediately commenced questioning whether this could in fact be happening.  Surely a conference couldn't sell out THAT fast?  Well, it did and realistically it was time to start planning for 2013 unless you were one of the lucky few who managed to place their name at appropriate altitude on the waiting list.  The problem with waiting lists is that they involve waiting and it was only last night after two months of finger-crossing that I discovered I was one of the few who would be allocated a place.

I was excited to get the news and immediately began to wonder just what the big stories would be this year.  With so many companies developing new sequencing technologies either silently (e.g. Oxford Nanopore) or vocally (e.g. Genia) whilst rivalries heat up between some of the biggest players (Illumina vs. Life) and expectations of new technologies from waning stars (454) increase, we're almost guaranteed some big news at AGBT.  Therefore the second piece of news I referred to at the beginning of this post seemed to come more than a month too soon.

In a scoop that would rival Clark Kent for journalistic rapidity, Nick Loman blogged that Ion Torrent had just announced a new desktop analyser that will launch this year.  This announcement paradoxically seemed simultaneously predictable and completely unexpected.  Anyone who has purchased or considered purchasing an Ion Torrent PGM will be aware of the projected scalability of the machine.  Most probably expect that it will some day live up to its acronym and achieve a level of throughput that actually enables Personal Genome Sequencing, however the announcement of the Ion Proton and its ability to sequence a $1000 genome in less than a day will have people wondering exactly what the plans for the PGM are.  As someone who has recently been considering purchase of one, I now have begun to take Ion's hype with an even larger dose of salt than I had previously.  Most people knew that 'the chip is the machine' meant scalability to a certain point but that eventually a ceiling would have to be hit.  I find myself wondering if this ceiling will now be self-imposed to ensure marketability of two products rather than occurring due to technology limitations.

I would love to get peoples thoughts on both points.  Are you coming to AGBT this year?  What do you think the big new will be?  Does the Ion Proton spell an end for the massive leaps in throughput for the PGM?  Or do both have the potential to develop alongside one another for years to come?

I look forward to your thoughts.