Wednesday, February 01, 2017

Illumina Drops NeoPrep

At the 2015 AGBT meeting, Illumina launched the NeoPrep, a ~$40K instrument to automate the preparation of up to 16 sequencing libraries at a time, using a technology called electrowetting microfludics. Now news comes that Illumina is dropping the NeoPrep, halting sales immediately and allowing existing users about a year of reagents.  What happened and how does it impact genomics?
Sequencing library preparation for Illumina generally involves a number of molecular biology steps, such as pipetting, bead cleanups, PCR and so on.  These require skill, and perhaps more importantly quite a bit of user time.  Very large laboratories automate many of these steps on large, expensive liquid handling robots.  NeoPrep was meant to enable mid-size operations to automate their library preparations with minimal hands-on time and requiring only a small number of pipetting steps at setup time.

NuGEN had previously commercialized a smaller but similar instrument called Mondrian.  The name was fitting, as the electrowetting technology involves moving droplets around an array of square pads; by altering the potential on the pads they can be made hydrophilic or hydrophobic.  Videos of such systems can be mesmerizing, recalling for me the low-res (128 x 128!) video games of my childhood.  Biochemistry meets Pong.

Illumina bought Advanced Liquid Logic in July 2013, from which NuGEN had licensed the underlying technology.  NuGEN then quietly shuttered Mondrian.  A year and a half after the acquisition, Illumina launched NeoPrep.  The shutdown notice just under two years later confirms rumors that the instruments had not been performing well. In their letter to customers, Illumina cited reliability issues in house and far worse ones in the field.  In addition to outright library failures, apparently sometimes the successful library would be physically unrecoverable from the cassette, slightly misplaced by the fluidics.  It can be inferred that manufacturing the expendable electrowetting cassettes to high standards is challenging, and perhaps the variations in buffers and contaminants with user DNA preparations simply make matters worse.

Apparently there were a number of happy NeoPrep users, as a number of disappointed tweets have appeared.
This would seem to leave Illumina users with no small-scale automation options.  Nextera preparations are perhaps the best choice now for small labs which wish a somewhat easier protocol, but Nextera libraries can give very uneven coverage and this is not appropriate for other library types, such as RNA-Seq.

Does Illumina see the failure of NeoPrep as creating a serious hole in their portfolio?  Firefly is supposed to have a library preparation module; if this is based on NeoPrep technology then that is a serious setback.  Alternatively, does Illumina have some other, presumably microfluidic but using something other than electrowetting, library preparation scheme under development?  

The perhaps even longer shadow is that Oxford Nanopore's VolTRAX library preparation device is also based on electrowetting technology.  Users have been tweeting pictures of VolTRAX devices, but not yet any images of preparing libraries with this or reports of running libraries.  This fits the pattern seen with MinION and PromethION, in which consumables for new ONT gadgets often greatly lag delivery of the devices themselves.  VolTRAX looks very different than NeoPrep, with a much smaller controller device for about $4K and the rest in the consumables (pricing not released).  If building the consumables consistently, or interference from sample impurities, are inherent liabilities of electrowetting, then VolTRAX may be in for a rocky ride.   Even if VolTRAX runs into the same wall as Mondrian and NeoPrep, ONT has another simplified library preparation gadget under development which does not use electrowetting, Zumbador.

AGBT is just around the corner; perhaps Illumina will clarify their strategy for the small laboratory market in that forum.  I'll be covering it remotely as I have in most years; as a clever character once said there is so much time and so little to do. Wait, reverse complement that...


Unknown said...

Thanks for this update Keith - I remember when NeoPrep first came out and all the complaints about only being able to run 16-at-time or else the $/sample goes way up, due to the monolithic 'all or none' nature of running that system.

Back in the SOLiD days Life Technologies has an automated library prep instrument based upon other technology Invitrogen licensed (I can't remember the name, looks like TMO replaced it with their Kingfisher technology on their website). It is an underserved market, either it's 96-well or hard to make economical. There is an interesting company that does automated Gilson Pipetteman though! (I looked it up - it's Andrew Alliance - very interesting, and not very expensive relatively speaking.)

The VolTRAX certainly does look interesting... BTW are you headed to AGBT this year?

Unknown said...

Thanks for the great blogs! However the letter actually says nothing at all about performance issues. Fake news? LOL