Wednesday, June 14, 2017

New Life in the Sanger Market

In my bit on "I'm not dead yet" technologies recently, I included large scale Sanger sequencing. That reflects to a large degree my personal experiences and biases.  Targeted Sanger is great for spot checking the occasional junction or misbehaving clone or strain, but I forget that many clinicians still see it as a gold standard.  Apparently there are others who disagree with me, as Thermo Fisher recently launched a new Sanger instrument targeted at small labs, and according to GenomeWeb Promega plans an instrument offering in the same space as well.
ThermoFisher's new SeqStudio instrument is intended to serve in small labs with light or erratic Sanger usage,   In that case, simplifying workflow is key.  So the new instrument features a click-in cartridge with all the consumables for running the reaction products out -- capillaries, gel and so forth.  Sanger reactions are still run outside the instrument.  In addition to Sanger reactions, the four capillary instrument can be used for fragment sizing in the same run.  That sentence is a bit hard to write, because Sanger sequencing is fragment sizing, so it must be a software thing because the larger instruments from Thermo are apparently locked into one mode or the other in each run.  The box runs $57K and apparently the sequencing runs will go for about $3.50 per lane.  Read lengths are spec-ed at up to about 600 bases, which is respectable but not spectacular.  

One thing definitely not helping my blind eye for Sanger is being in the Boston area.  At least two different companies offer not only next-day results but will send a courier over to pick up your samples.  If you miss the deadline, drop-off is also an option -- or if you like a coffee shop near the Sanger lab, which one past colleague of mine used as a reason to run the samples himself.  But such services aren't found except around major research hubs, so having on-site Sanger could be very attractive if you run it enough.  

Given this instrument, and perhaps another one from Promega, is the Sanger market really growing or just there are enough holdouts to maintain a healthy sales position for a company or two?  A bit of me hopes the market is seen as growing, as perhaps that would revive what once was a hot topic in journals: fully minaturized microfluidic Sanger.  In theory the entire chemistry (including thermocycling) and detection could be run on a single chip.  That could enable a much smaller device.  But nobody is going to invest in pushing these from laboratory demonstrations to field instruments without a promise of making some serious bucks.

And therein lies the problem. Certainly there are still a lot of people who trust Sanger or who have assays particularly well suited for Sanger.  If you want to collect a small amount of data from a small number of samples or even a small amount of data from quite a few samples, Sanger may still make economic sense.  But there will be a constant pressure coming from increased use of gene panels running on different boxes.  With devices like MinION, with its washable flowcells or small, cheap "crumpet" SmidgION/Flongle flowcells, or perhaps the Roche/Genia technology, the cost per sequencing run could drop substantially.  There's always the quality issue, but ONT makes regular progress there and Genia looked very interesting.  There's also the BioRad/GnuBIO instrument which may drop on the market soon.  The point is that the cost of the competition may be dropping so that even a very small gene panel on a small number of samples may be cost effective to run on newer technologies.

There's also companies like seqWell, and now iGenomX, bringing the cost of shotgun sequencing clones down.  That's another market remaining for Sanger, as sequencing small clones can still be more cost-effective on Sanger than other instruments, particularly if you have a very large number of relatively small clones.  But if library construction can be reliably done en mass for single digit  dollars per library, then Sanger may be nudged out here as well.

By the way, on that idea of Sanger as gold standard? To me it is too often a pyrite standard, as in the high G+C world I live in compressions can still be problematic.  Indeed, there is a paper in the last few years which repeatedly explains away a missing key residue in a protein known to be functional.  Alas, the residue is only missing in the ancient Sanger-derived sequence in GenBank; do a proper job and it is there.  I used to be unable for company stealth reasons to admit a deep understanding of Streptomyces rapamycinicus RapK, but that lid has been well blown. Starbase has built a whole technology around "rapalogs", synthetic molecules borrowing a core structure from rapamycin, so it shouldn't be surprising I've stared at that cluster a lot.  Plus there's another rapamycin cluster now in the database that gets that sequence right.

Given inertia and market forces, Sanger sequencing may well remain relevant for at least another decade.  It's already been over 40 years since Fred Sanger's genius came up with the idea of terminator nucleotides for sequencing.  Terminators also form a key concept in the Illumina, Helicos and QIAGEN sequencing platforms as well.  Some day Sanger sequencing will be retired, but that day isn't anytime soon.


  1. In my view Sanger sequencing serves a very different purpose than new sequencing technologies for at least my big chunk of the research field. I don't even think of our Illumina/PacBio work as being simlar to anything we do with Sanger, although they are all sequencing. We spend a lot of time constructing many smallish plasmids for essentially everything our genetics lab does, but we will make these sporadically, 1-2 this week, 5-6 the week after that, maybe 50 in one day next month. Any time these get made we need to verify that they were constructed correctly, and this can be done sufficiently with 2-3 Sanger runs per plasmid. I'm also spoiled by three different commercial companies with daily dropboxes with next-day service less than 200 feet from my lab.

    I'd love to know if there were any newer technologies remotely competitive with this (or the ABI 3730 downstairs). It would be nice to regularly get back full plasmid sequences instead of just our area of interest, but doing a library prep on individual plasmids even using the most economic methods I can find is already twice the cost of our Sanger runs, not including paying someone to sequence them and waiting for that to actually happen, unless we were doing dozens all at once in which case manually setting up a whole plate of Sanger reactions becomes worthwhile and even cheaper.

    Are there any newer technologies/services that let you get back sequences of small numbers of small plasmids or fragments economically and in a reasonable timescale?

  2. Jonathan:

    Thanks! What cost per plasmid would tempt you to switch from Sanger?

    I'm guessing there isn't a cheap enough technology to best Sanger for your application, but useful to know the exact parameters.

  3. The MinION is remotely competitive, particularly if you're sequencing lots of sequences of different lengths (or sequences that are easily separable in post). Consider that 12 samples with 10 different 1-10kb amplicons per sample will have a marginal reagent cost of about $6 per sequence (after the DNA is extracted/amplified), and that's assuming the flow cell is not washed and re-used.

    Consensus quality is an issue for MinION, but cost doesn't need to be.

  4. It is hard to beat Sanger on low throughput low cost sequencing. Library construction costs are a concern for any other method. We do have a service that offers 2nd gen sequencing for $25/sample (Univeristy price) with enough coverage for plasmid size samples. Some people have pushed this up to BAC-size (120 Kbase). However being 2nd gen (miSeq/hiSeq) there is assembly involved which, of course, can be problematic. We have had some very nice results and then some poorer results.

  5. I wonder if it has to do with a lack of awareness of the sequencing kits that are out there.

    A Miseq Nano 300 cycle kit runs for ~$240 and ~$290 for a 500 cycle kit, including our institutional discount. This gives you about a million reads. You can fragment at least 10 plasmids with fragmentase or even plain old nonunique restriction enzymes and get a really high quality sequence assembly in less than a day. You can do the same for sequencing larger amplicons.

  6. That's bare run costs, and the library prep costs are in the area for 30-100$ per sample...

    Also don't forget bioinformatics costs (not all users know how to prepare a bam file from their custom vector reference and run reads and call the variants).

    Also prices in the Europe/UK are a lot higher for most users of the Illumina MiSeq kits.