Tuesday, July 12, 2011

Ion's Gaggle of New Accessory Products

     Ion Torrent had a press release announcing a number of new components to the platform, as well as shipping of the 316 chip and a drop in price of the original 314 chip to $99 (alas, the announcement was not synchronized with updating the online store, so no details could be checked -- such as the price of the 316 chips).
     A big piece of the announcement is the 316 chip release along with the price cut on the 314.  This fulfills the promise of continual jumps in performance, and I'm personally itching to try out the new chips on several projects.  Dropping the price of the 314 chip to $99 will be great for labs that want to experiment, particularly those gunning for grand challenges.  The 316 specifications promise 10X more reads but from only 6X more sensors; process improvements account for the rest.  Those same improvements should apply to the 314.  As noted here previously, consistency in number of reads per run has been an issue in the past, and it isn't clear whether this has been addressed.
     One somewhat contestable claim made in the press release, and extended in an online self-back-patting video, is that Ion now has the fastest DNA to data process in the business, stated at 9 hours.  The first contestable issue is that this requires using the unreleased (but promised this quarter) Ion OneTouch template preparation system.  The OneTouch has apparently now acquired a new capability of enriching for templated beads; it was previously ambiguous whether the OneTouch performed any enrichment.   It also isn't clear where in the process that library quantification fits in; a new qPCR kit is announced for this step. 
    So Ion is claiming 9 hours for a 100bp run and 8 hours for a "shorter" run.  However, to compare unreleased product to unreleased product, it would seem the MiSeq with Nextera library prep could probably beat this number.  Similarly, PacBio might contest here with their simple no-PCR library prep, though they have not been aggressive about publicizing their process time.  Roche/454 continues to appear to cede this space to Ion by default, an attitude which will hurt them dearly if Ion can push read lengths up.  
     On quality, a claim of order-of-magnitude improvement over the original system is claimed, but no more detail given.  A promise of 200bp read lengths this year and 400bp read lengths next year is made.  New software cuts analysis time from 60 to 30 minutes.  
    Rounding out the announcement, a 16 barcode kit was announced as an early access item, a new fragment library kit with enzymatic fragmentation and a new control kit.  The barcode kit is a nice complement to the 316 chip, as with a 1M read run there are many more projects that could benefit from barcoding vs. the 100K reads off a 314 (though certainly some projects could barcode there as well).    However, it is early access and 16 barcodes is a rather limited set.  I can't examine the new fragment library protocol; it's hard to believe it is as easy as Nextera, but if Ion can make things simpler and obviate the need for a shearing instrument those are wins.
    All of these new products appear to be useful additions to the platform, but if Ion really wants to succeed I believe they need to lasso some other companies to adapt their kits to the Ion universe.  While Life might want to hoard all the revenue from the platform, the Illumina world in particular has had the benefit of many manufacturers pushing the envelope in different areas, giving scientists choice and also kits optimized for different purposes.  
     With increasing throughput, Ion will really need to get some targeted sequencing kits in place.  While right now the machine is well suited for PCR-based methods, eventually it will have the capacity to make hybridization selection feasible.  Even in the PCR field, I haven't heard any noises yet about any platform other than Fluidigm.


Anonymous said...

to compare unreleased to partly released the Miseq lists 19 hrs to do amplification and sequencing so given that amplification will take the same time (lets say an hour as cbot is 4 hrs for much larger chip) it seems safe to estimate that 100 bases on a miseq is 1hr for amplification + 9 hrs to do run so 10 hrs plus have to do library prep and analysis, According to Ion its 8hrs and if add on 1 hr to do DNA prep then do the whole thing in 10hrs plus the sequencer can be used multiple times in the 10hrs the miser is tied up .

Anonymous said...

I think they are planning on releasing the TargetSeq capture kits for SoLiD and PGM. The PGM details is non-existent though.