OVERVIEW
On January 4, we spoke with a program director at a large biomedical and genomic research center. We came away from the discussion with a greater appreciation for the recovery trajectory during this [3rd and final?] wave of the coronavirus pandemic, as well as how entrenched Illumina (ILMN) is as the leader/benchmark in the industry.
HIGHLIGHTS
- Most labs are either at or back to within reach of full capacity; however, if priorities shifted toward SARS-CoV-2/COVID-19 testing, priorities may have shifted. For our contact at a large academic lab, the expectation is for cancer sequencing activity to be back to 90%-100% by the end of March (from ~75% today), but the lab is otherwise fully operational as of early January.
- Feedback on NIH grants suggests that our view of the backlog is correct: the pile of money, which has been stacking up, is burning a hole in labs’ pockets and is likely to be spent as quickly as possible before the September fiscal year end.
- Positive feedback on Twist (TWST), which appears to be uniquely positioned to be helpful to labs -> we’ll keep a closer eye on it.
- Feedback on MRD was mixed, but if we are right about Natera’s 1) moat being defensible, and 2) favorable cost structure, then the takeaways are at least confirmatory of our positive outlook for SigNatera, in our opinion.
CALL NOTES
What happened w/ your lab during the initial height of COVID-19 and how open are you now?
-
Things were totally shut down from mid-March through mid-July, so there was no sequencing at all. It was a combination of pandemic response and spinning up SARS-CoV-2 vs. genomic.
-
We now have several [business] units to balance and COVID diagnostics have increased in priority.
-
80%-90% of the data generated through the lab is a result of Illumina standard whole genome sequencing.
-
Cancer sequencing - smaller, complex, targeted whole genomes, whole exomes, single cell, etc. - are running at ~75% capacity. I’d expect to be at full capacity early this year (90%-100% by the end of March). The lab is fully operational w/ staff in the building - everyone gets tested [for COVID-19} every day, and the COVID testing is running at close to 100k tests/day. That should become more predictable and decline a bit, but we’ll see.
Have you noticed any other issues/disruptions as a result of COVID-19?
-
There have been some supply chain issues for all the products needed to run testing and sequencing. One of the bigger issues is plastics out of China -> pipette tips, for example, and then consumables are in high demand.
-
With pipette tips, those are needed for the robotic systems. Every time you switch to a different vendor, there’s validation because labs are running clinical programs under quality control. You can’t easily swap out a product like that - it’s not just the grade of plastic, there’s a specific catalog number, etc.
We see [NIH] grant volume trending up for things like single cell and there’s a backlog… if you’re awarded money, what happens if it just sits around? It looks like the cash balance is growing.
-
If you get a grant, you project the budget per year and then can get the spending for that amount per year. You must account for it. If the money isn’t spent in the specified period, it can be clawed back.
-
Grants typically decrease over time - a little less money each year. So, there’s high motivation to spend everything every year. Having COVID-19 shut down labs is a big deal. For the most part, a global pandemic is a reason to carry an amount over, but sometimes there’s pre-spending on reagents and such.
-
With sequencing, you’re making a bet/projection based on estimated costs - e.g., that the cost will be 10% less next year because Illumina is more efficient or you’ve put processes in place to make it cheaper. It’s up to labs to realize those and while there’s a low risk that funding is taken away this year, if projects are mismanaged in the wake of the pandemic, funds could be lost.
-
Contract-based work is different - labs often do work through third-parties, not grant-based, and generate data like a commodity.
Given competition (i.e., BGI) and pricing dynamics, do you think Illumina’s volume can outstrip price deflation?
-
It’s true that the “best days” are in the past, and I think samples out there to be sequenced are finite. That is, in terms of samples that are ready, prepped, and available, there’s a limit - it’s why there isn’t infinite growth there.
-
In the long game, when sequencing is something everyone does - for 23andMe or other consumer-based testing (not just medical/research) - it could be different. I think that’s why interest in Grail is so high. Early detection of cancer is interesting. I want the test - I’ve been impressed by the data.
-
I know the jury is still out on early detection. If there’s broad deployment, will the community be able to interpret the results? Also, I think it’s too early to say whether Grail or Thrive has the most effective technology. I am a little worried that it’s so complex to interpret... What will the follow-up look like? It could be a slow adoption for all of them. What clinical tests or action does a result drive? What will insurance pay for (big question)? How many people get it to follow-up?
Do you have a view of 10x Genomics and NanoString?
-
Spatial transcriptomic protocols - to see if it can help w/ generating the data - has a lot of interest right now. The ability to pinpoint activity spatially on tissue helps to paint the biological story from all the other data generated. The more samples we do through current tech, we can develop better analysis tools.
-
I get asked about their scalability often - they are the new “bright, shiny objects.” It’s all about whole genome, single cell, plus spatial. I’m more familiar with 10x, but NanoString is popular with custom RNA too.
Twist Biosciences
-
TWST is a source of panels - they are a great reagent development partner (for targeted cancer panels). Before Twist, there was no good option for gene lists to target for high quality, balanced design.
-
Helpful with specifics - balancing gene panels and cost competitive way to get small amounts of reagent (e.g., if you only need enough to do 100 not 100k samples). With Illumina, you couldn’t get less than 1k samples worth of reagent.
-
Illumina used to have a major targeting reagents business, but that’s not the main business anymore. Twist exome has taken share and has momentum - it’s an easier workflow (library construction w/ capture kits, easier to automate).
-
Allows for work on one gene or other genomic territory that’s interesting for a rare cancer - not on a predesigned panel (it’s too expensive to get a custom panel).
-
TWST has very responsive reps, they are very smart on design, targeting, etc.
What do you think of the growing role of technology, computational biology, data problems?
-
Larger projects generate tons of genome data that people want to search and navigate. The data needs to be accessible by a typical researcher at Stanford, for example. You can’t download everything, store it, etc. (would cost $100s of thousands).
-
Data must be put somewhere in a pre-analyzed way w/ phenotypes identified/tagged: high quality vs. low quality geotypes, what parts of genotype are not coerced...All those are important questions.
-
There are more “data science” grants, yes.
-
New hardware is being developed, including specialized processors and other tech hardware.
Thoughts on molecular residual disease (MRD)?
-
MRD in general is a highly competitive space. It comes down to a cost question. It’s hard to do enough sequencing to get sensitive detection - i.e., Do we know the test is accurate?
-
There can be lots of false positives w/ sequencing - looking for something at 0.1% of reads - whole things lights up - requires significant adjustments and error correction behind the scenes to know what’s a positive hit.
-
Many researchers are skeptical of how to effectively track MRD - in which cancer(s) is it most appropriate? Where can it make a real clinical impact? Like early detection - does it change the outcome?
-
With Twist, you can have a panel for specific variants - a personalized cancer panel. Sample that patient for cfDNA to see if variants light up again in blood. But, the cost element is NOT favorable enough. Going forward, if this is the kind of cancer care we want to offer (as a system), patients should go in every 3 mos to check cfDNA and it must be cheaper and more specific (less of a black box than some of the droplet PCR).
NIH Grant Update | January 2020
Please reach out to with any feedback or inquiries, questions for future field checks, or requests for underlying data.
Thomas Tobin
Managing Director
Twitter
LinkedIn
