While Oxford Nanopore and Ion Torrent had the most people talking at ASHG (see ASHG Sequencing Platform Announcements, part 1), Illumina and others used the venue to make some significant announcements as well.
A little over two years ago Illumina upgraded their HiSeq platform to generate 600Gb of output per run, tripling the output of the original launch specs of 200Gb. Initially this actually hurt Illumina a bit financially with customers electing to save money rather than filling up the newfound capacity with new samples. But after a couple of quarters their users found out how to fill those monster runs and Illumina cemented itself as the undisputed leader in the high throughput segment. Not so coincidentally, a little over two years ago is when the previously precipitous drop in sequencing costs started to flatten out.
When the new 600Gb chemistry was announced Illumina also stated that they had been able to achieve internal runs of greater than 1Tb (using longer read chemistry). But as the focus turned to the new MiSeq platform, we never heard anything more about 1Tb runs. Until now. Illumina announced at ASHG a new kit for 1Tb runs will be available in early 2014. Even better news, most of the gains now come from increased density (2X125b reads across four billion clusters), so the runs are only 6 days. Now the bad news. This upgrade is only available for factory-built HiSeq 1500s and 2500s and HiSeq 1000s and 2000s shipped in 2013. If you have an older HiSeq (even it was upgraded to a 1500 or 2500), you can’t upgrade to the new higher outputs. This upgrade probably requires some new optics and hardware that would be just too expensive and involved to upgrade on the older machines. Still, 600Gb isn’t too shabby. How exactly this upgrade will fit in with Illumina’s previously announced future improvements, 2X250 runs and flow cells with semi-ordered clusters, is unclear. Clearly this workhorse of a platform still has some headway.
PacBio has been making steady improvements to the RSII, increasing outputs and read lengths, winning a lot of converts along the way. The problem is they have a very expensive machine which, while quite useful, currently only occupies a niche in the next generation sequencing market. What would help, and what many of us have been waiting for, is a smaller, less expensive instrument. And it appears that this is exactly what PacBio is working on. This didn’t seem to make a big splash at ASHG, but that’s probably because Roche talked about it in one of their workshops. No offense, but these days probably very few people who are keenly interested in next generation sequencing would bother to attend a Roche workshop when so many other great talks and workshops were on offer.
Finally, BioNano Genomics announced a version 2 chip for their Irys System. The V2 has 15 times the output of the original chip, yielding 50-100Gb of output per run. This is enough to make genome mapping feasible with human samples, and really ramps up the pressure on OpGen and Nabsys (and possibly Oxford Nanopore)