Most sequencing platform companies made their big announcements ahead of AGBT, including MGI. However, MGI hinted that there would be more to come in 2023, and it turns out AGBT is recapturing at least a little bit of its former glory of being THE place to make a big announcement. MGI, under the marketing efforts of their US subsidiary Complete Genomics, launched three new products: DNBSEQ-G99 (a low throughput sequencer), DNBSEQ-T20 (discussed below), and stLFR Library Prep Kit (single tube long fragment read assay). The T20 seems to be the most groundbreaking of the three. It’s a behemoth that can sequence 50k genomes per year at $99 each. This is their “deconstructed sequencer” that includes open slides (as opposed to flow cells) and a robotic arm that moves the slides between reagent baths and to/from the imagers. By converting to this more open arrangement they can maximize the feature density, decrease reagent volumes and thereby reduce the overall cost per Gb. The system includes a robotic arm and two imagers which utilize six slides per run. It generates up to 22Tb per day and 72Tb/run (implying a run time of a little over three days).
So let’s think a little more about the pricing. A $99 genome sounds great, but what about the instrument cost? Well, it turns out CG isn’t really selling this instrument directly, but rather via a reagent rental model. So the $99 includes the instrument amortization. Given that, now we need to know over what period of time the instrument is being amortized. This is especially important as this price requires the pre-purchase (or at least commitment) of enough reagents for 50k genomes per year. Unfortunately, CG isn’t being too forthcoming. All they said to my question was “less than 5 years”. Given that response I think it’s reasonable to run some comparisons at 3 years (150k genomes) and 4 years (200k genomes). For comparison, Illumina’s NovaSeq X Plus generates 20k genomes per year at an amortized rate of ~$221 over 3 years or ~$216 over 4 years. The only uncertainty I have about this being an “apples to apples” comparison is the compute and maintenance costs. Illumina includes the basic compute needed, but I’m not sure if the T20 does. Illumina maintenance contracts are typically ~10% of the instrument (which would raise the per genome cost to ~$227 for 3 years or ~$222 for 4 years). I don’t know if the T20 $99 price includes maintenance and, if not, what a typical maintenance contract would cost.
So what does this mean for the market? Regardless of the minor details around pricing, it’s pretty clear that MGI has set the bar for low cost, ultra high throughput genomes. As I outlined previously, Element seems to have the lowest price when sequencing ~3k genomes per year, while Illumina has the lowest price when sequencing between 5k and 20k genomes per year. MGI’s new DNBSEQ-T20 now means customers have an even cheaper solution starting at ~20k genomes per year (with 50k genomes required to get down to the stated target of $99/genome).
This new ultra high throughput sequencer won’t be available for ordering until Q3 of 2023 (and no word yet on delivery dates, but my guess is they would bend over backwards to get at least the first shipments completed by the end of the year). Given that long wait, they’ve taken a page out of Element’s book by offering a price of $1.3/Gb for customers willing to purchase three DNBSEQ-T7s now. Just a little something to tide everyone over until the real thing comes along later this year.
With the launch of the DNBSEQ-T20, the market has a new low water mark for ultra high throughput genome projects. This is the latest example of new competition really opening up the market with customers having more choice than ever, with Ultima, Illumina, Element, and MGI now all with new low cost options.