454 Junior Announced - A New Strategy for Roche?
20 Nov 2009
A press release from 454 was brought to my attention via their Twitter feed yesterday. We'd heard rumours about this already, but Roche have confirmed that they will be supplying a new bench-top sequencing, the "GS Junior" in 2010. So other than a cute name, what will this bring to the party?
"The GS Junior System, which is no bigger than a typical laser printer, has performance and features tuned to fit the needs of small to medium sized laboratories. The platform will launch with long-read GS Junior Titanium chemistry, offering 400 – 500 bp read lengths, and will be shipped with a desktop computer that is optimized for GS Junior Run processing and downstream data analysis."
A laser printer form factor sounds attractive, but this is offering daisy-wheel performance with a throughput of 35 million bases (c.f. 500 million from the GS FLX and 30Gb from Illumina) per run.
Contrast this strategy to the other big players in the high-throughput game, Illumina and ABI. Their strategy for the last few years has concentrated on inexorably increasing throughput with an eye on the glittering prize of human genetics - the $1000 and then $100 human genome.
Roche are going in quite a difficult direction and shooting for a different market. ABI still dominate sequencing for smaller customers (i.e. non-dedicated genome centres) such as the molecular biology or clinical diagnostics lab with the wildly successful 3700 series. It can be seen from our map of high-throughput sequencers that whilst Illumina are concentrating on volumes and genome centres (80 machines at the Broad, 40 at Sanger), 454 have slowly managed to get small volumes into larger numbers of institutions. Having said that, the machines are typically housed in sequencing facilities if not sequencing centres. They don't yet sit alongside the battered thermal cycler and pipettes in the average lab.
I can see Roche having some success with this strategy. But it will require a change in mindset from users. PhD candidates sequencing individual gene loci or plasmids in their mutants will start switching to whole-genome resequencing approaches (did my site-directed mutagenesis experiment affect other parts of the genome?). Clinical diagnostics labs will start looking at tens or hundreds of gene loci (say, all genes known to affect cardiovascular risk). Clinical microbiology labs will routinely deep-sequence viral genomes to help treat infected patients. They may even start whole-genome sequencing bacterial isolates for pathogen profiling and help with infection control.
The Roche Junior may be the perfect solution for these new approaches which are surely to come.
But I see two major problems right now that need sorting for this to happen.
1. Speed up and simplify the 454 library construction
It takes us days to weeks (for paired-end libraries) to prepare a 454 Titanium library for sequencing. There are over 20 steps, some of the molecular biology is quite hairy and if you are not careful, you can easily lose your library and have to start again. You need a couple of weeks training before you can touch the system. This won't be acceptable for the average grad student or post-doc. The website states that the sample prep will come down to just half a day, which sounds very promising.
2. Plates and reagents for 454 are very expensive
You are still looking at about £6000 ($10,000 US) minimum in consumables to run a single plate of 454 sequence and get about a million reads back. How much will the Junior's consumables cost? I'd estimate they'd need to get this down to way below £1000 a run for this to be adopted in small to medium labs enthusiastically for new applications.
Also buried in the press release is some more detail about the updates to the FLX Titanium kits. As we'd already heard, reads of up to 1000 base-pairs are promised. This will bring immediate benefits to a number of applications, particularly amplicons and transcriptomics. We will certainly be using to for phylogenetic profiling of bacteria, using the extra long reads to capture more 16S variable region and hopefully improve the accuracy of these approaches.
The press release also hints that less input DNA will be required to operate the new kits. This is great news. we are currently asking our customers for between 5 and 10 micrograms of DNA for input which is sometimes difficult to achieve. They don't specify how much is required, but certainly Illumina are winning on this, requiring much less than a microgram as input for their library construction.
Read the full press release.
Update 14:08 GMT
The guys from 454 (cheers Tim!) have been in touch to give some indicative pricing for the machine.
Pricing information removed.
So there you have it, less than 15 years after the first bacterial genome sequence was completed (the biological equivalent of an Apollo landing journey back then) an average lab can do the equivalent for £x,xxx on a machine no bigger than a laser printer in a single day.
And on the subject of DNA requirements for the new FLX protocols, this has been cut down to a minimum DNA template requirement of 500ng (thanks: genejockey).
Update 15:21 GMT
Woops, seems like 454 guys don't in fact want the pricing released to the public just yet. Hopefully they will announce in the near future but have taken it down at their request.