The breakeven price for genome sequencing

A brief history and predictions

Our bodies have our innate blueprint, our genome, repeated 30 trillion times. In the whole course of life of Earth, we’ve only known about DNA since the the 1860’s when Friedrich Miescher isolated it from white blood cells. The structure was famously solved by Watson and Crick in 1953 with the Human Genome Project finishing a 13-year quest for a human genome sequence in 2003. Since then the cost of sequencing DNA has dropped precipitously allowing more information regarding disease dispositions, heritage, physical phenotypes to be decoded.

DNA sequencing technology has been so commoditized that, for the past 16 years, laypersons have been able to buy kits to sequence parts of their genome. 23andMe started shipping its kits in 2007 at a price of $999. 23andMe and its successors AncestryDNA, MyHeritage, Family Tree DNA have continued to provide consumers insights on their traits, with other companies such as Reveal enabling paternity testing and Embark, Wisdom Panel, and Chewy even offering trait insights for animals. Some companies such as Nebula, Centogene, and Sequencing offering more complete tests of sequence and analysis of your entire genomes instead of SNPs (single nucleotide polymorphisms, aka point variations), which most incumbent tests offer. Using these tests, some 30M people have sequenced at least some part of their genome.

We know that genome sequencing insights such as those above are a neat (and sometimes medically helpful), novelty product. In aggregate, however, genome sequences matched with disease phenotypes are tremendously valuable for mechanistic understanding of disease which can be used for better treatments, clinical trials, and ultimately, cures. After all, genomic stratifications of patients in clinical trials doubles the chances of eventual drug approvals. Given that clinical trials are the primary expense in drug discovery companies, there have been pushes to access larger pools of genomics data. GSK was a leader here and commenced a large, $350M partnership to 23&Me for access to ~9.6M people’s DNA sequences. A calculation puts the cost per partial genome (likely a single nucleotide polymorphism, SNP, array) at ~$39. This is where things get interesting since the market rate of $36 per genome is only one order of magnitude less than the cost to sequence a genome today ($399-$999). And, if we’ve learned anything in the past 20 years after sequencing the human genome, it's that sequencing costs drop exponentially over time. This is mainly due to advances in next generation sequencing or even more basically, putting biology on chips. Sarah at Rough Diamonds points out that most things (DNA sequencing, solar panel productivity, medical imaging) put on a chip follows exponential cost and efficacy curves. 

With this lens, I extrapolated the trendline of the cost of sequencing a human genome. This was a thought experiment and not meant as dogmatic eventuality. However, I did find it scintillating that if costs of sequencing were to continue decreasing the cost to sequence a human genome might be $1 in 2027. (As a caveat, Ultima’s $100 genome results weren’t included in the dataset from the NIH and given the controversy in sequencing depth weren’t included in the analysis below*). 

The graph above garnered some fair pushback that I hadn’t taken into account the decay rate seen in the cost of sequencing in recent years. To address this I did a separate cost curve of the past 10 years of DNA sequencing. When doing an extrapolation here, the $1 genome is projected to be realized in 2041. In this cost curve of the $10 genome in 2033 which still undercuts what an individual's SNP array is worth to pharma at $35+ as per the GSK and 23andMe deal. In terms of the breakeven point between cost to sequence vs. market rate, I’ve heard from folks in the industry that whole genome sequences can be sold to pharma at $250 each. The $250 price point is already more than what Ultima is claiming its whole genome sequencing cost to be. I’m obsessed with the breakeven point because it’s the start of a changing incentivization structure, where people might be able to profit from getting their genome sequenced.

The new cost incentivization structure could lead to many more healthy adults sequencing their genomes and brokering its sale to companies as a second revenue stream. This is neat because its encouraging a different, and more diverse group of individuals to get their genome sequenced. Currently, the genome data collected is skewed towards wealthier populations, particularly those in Europe in North America. However the low cost of genomic data in the near future unlocks more customers and more genomes for science.

Companies such as Nebula Genomics, Encrypgen, and Genos have all predicted an eventual marketplace where people have autonomy and control over the sharing and the upside of their genomic data. However, they have been working in times where sequencing costs are largely still more expensive than what people are willing to pay for it. The window to the changing dynamics are closing sharply and we think companies building to the future of the $1-100 genome will be particularly well-placed to contribute to bio marketplaces which enable the easy and secure brokering of genomic data. Cheap sequencing of the human genome leads to different technical problems which will be required to bring down the total cost of the test(including shipping, DNA extraction, and lab automation). 

Genomes are only the tip of the iceberg when it comes to biological data collection and limitingly don’t give a functional understanding of what’s happening in the body. The layer above DNA is RNA, which can be sequenced using similar processes and gives functional readouts of genome regions which are actually expressing. The technologies involved decreasing the cost of DNA are also decreasing the cost of RNA. Beyond that, there are strides being made in other important biomarkers such as protein sequencing with Glyphic Bio, Portal, Nautilus. Metabolome deconvolution is correspondingly increasing in sophistication which is certainly a space to watch (Enveda Bio is a company to watch here). Although these are still more expensive than DNA and RNA sequencing we see a future where all biomarkers will be able to be collected, stored, and shared, to the benefit of the individual. Ideally this data will be layered with electronic health records (EHR) and be paired with increasingly sophisticated wearables data (Apple Watch, Levels, etc). What this is eventually building towards is a clinically relevant and monetizable experimental digital twins for each of us. Although a seemingly distant future, companies such as Unlearn AI are already making digital twins of people to help in the clinical trial process. Just as DNA sequencing has been commoditized to the consumer industry, so will this multimodal tech which will provide actionable feedback for folks to live healthier, happier lives. Not only this, but people will be able to profit from sharing all the data they’re collecting, leading to many virtuous feedback loops.

Taking a step back, the panacea of a future we’re building towards starts with a breakeven price for a human genome. Life science innovations matter and, many times are holding the floodgates for entire subsectors to be realized. I deeply understand this to be true with both DNA, RNA, and protein sequencing and synthesis technologies. If you’re building here, in any capacity, don’t hesitate to reach out!

*The proposed $100 genome by Ultima Genomics was developed by engineering the rectangular flow cell into a circular wafer flow cell which is a low-cost semiconductor material, increased surface area allows for more sequencing by synthesis reactions to occur simultaneously, the reagent dispensing is reduced by the inertia-driven spin dispense system. This allows for time increases in reagent delivery and imaging of the DNA sequence itself. Paper described here - https://www.biorxiv.org/content/10.1101/2022.05.29.493900v4