Blog

The current state of genome sequencing hardware and reagent costs, modified from https://twitter.com/LAbizar/st...

Keith Robison has a nice blog post on the Ultima technology, along with some discussion of the path to future drops in sequencing cost, down to $10 per genome or even lower. In that post, a sentiment attributed to Vinod Kholsa caught my eye: “when it costs more to FedEx the sample than to generate data one might think about not pushing farther”.

Effectively we are rapidly approaching a world where sequencing costs per se are irrelevant to companies looking to implement genetic testing at massive scale. The scientific implications of this are fun to consider, but it’s also worth thinking about the overall business implications. Specifically, if a billion bases of genome sequence will cost pennies in the not-so-distant future, we’re drawing nearer to a place where the ‘too cheap to meter’ paradigm might apply – the analogy to sequencing is the data in your cellular plan (now unlimited in some cases) or WiFi plan.

So getting out a crystal ball, what will the genomics industry look like in 2034?

1. Hardware companies will still be big, but not the biggest companies in genomics. Right now Illumina is the largest genomics company by far (as of today a $23B market cap), while specific genomic applications like diagnostics companies Myriad and Natera ($2B and $8B, respectively) are significantly smaller. In a world of commodity sequencing data, this will be reversed the same way Facebook or Google are more valuable companies than Comcast or AT&T. Incidentally this was probably part of the thesis for Illumina’s ill-fated acquisition of Grail – an important sequencing-based application company will be a better business proposition than a sequencing technology company alone.

2. We’ll see subscription models on sequencers. Currently, anyone who buys a sequencer pays 1) an up-front capital investment into the sequencer itself, 2) an annual flat service contract, and 3) per-run reagents. If costs continue to decline but buying reagents causes a user to think twice before doing a sequencing run, why not just roll reagents into the service contract? It might be more profitable to encourage as much usage as possible with a subscription, which in turn could drive purchases of more new sequencers and the corresponding subscriptions. In some way this isn’t too far from the Element $200 genome offer, which could be thought of as a $500k/year subscription.

3. More companies have sequencing as a non-’core’ part of their business. Since software famously ate the world, it’s difficult to imagine a company that operates without extensive software/IT infrastructure. Is your core business drilling for oil? You have an IT department. Constructing skyscrapers? Definitely have an IT department. Analogously any industry with even a hint of biology will be implementing sequencing. At Gencove we are extremely bullish on sequencing in agriculture and drug development, but as sequencing costs in themselves become negligible even industries well outside of traditional biotech will have sequencing departments (even if you’re drilling for oil? Maybe)

4. The first universal applications of sequencing will be unlocked by creative business models. Ever since the cost of sequencing began its steep decline, scientists have recognized that the end state of sequencing in different industries is simply ‘sequence everything’ – agricultural companies will sequence every animal that contributes to the food supply, hospitals will sequence every newborn, pharma companies will sequence every clinical trial participant, and so on. The simple reason this hasn’t happened yet is that in most (if not all) industries these sorts of ‘sequence everything’ financial models don’t pencil. Even over the past year, we’ve seen this change in conversations with customers – drops in sequencing costs are now allowing for experimentation in business models that will enable some of these applications to finally reach fruition.

Overall, it’s probably fair to say that the sequencing industry hasn’t been this dynamic since the late ‘00s and early ‘10s. At Gencove we’ve certainly seen the bottlenecks in genomic workflows shift from data generation to management and analysis of data at scale; this suggests more companies are starting to think about the world where sequencing itself is not a bottleneck at all. Are there other likely outcomes of this that I’ve missed? Let me know at @joe_pickrell or get in touch with us at Gencove.