My Year Old Advice To PacBio
I posted this almost exactly a year ago after their Q1 revenues and massive drop in stock price. As we’re discussing recent PacBio layoffs and would should have been done on the Discord, I thought it might be interesting to repost it publicly.
As mentioned in my previous post with PacBio releasing the Vega and (maybe?) dropping short read platforms they are now more or less going along with this advice…
But as we’ve been discussing on the Discord… the “Vega” class instrument perhaps could have been developed less cost intensively, and I would have dropped the short read stuff much earlier.
Anyway… have a read.
…WHAT FOLLOWS IS ONE YEAR OLD POST…
PacBio released preliminary Q1 revenue and then the stock price dropped ~50%. Market cap currently stands at ~$375M:
They’ve also said they are going to reduce burn with one analyst quoted on GenomeWeb saying "PacBio commented to us that it may reduce its headcount by triple digits to generate these expense reductions.". Another commented on influencing factors:
"[Macroeconomics] seemed to be the biggest culprit in driving the miss as funding remains tight for genomics across the board and was exacerbated by an inherently longer sales cycle associated with new product launches,"
Well That Sucks!
This is kind of a shame, because while they’ve made some dubious decisions. PacBio seem to have executed technically really well with the Revio. But PacBio have a lot more going on than just the Revio, at last count the following instruments are in development or released:
Onso: A NextSeq 2000 class mid-range short read sequencer
SuperOnso: A NovaSeq X class high-throughput short read sequencer
RevioMini: A Benchtop long read sequencer
Revio: The current-gen long read platform
SuperRevio: Some kind of undisclosed high-throughput long-read sequencer
In addition to this the Sequel II is still a supported platform.
So PacBio are spread pretty thin. It would be pretty easy to argue that the Omniome and Apton acquisitions were a mistake at this point.
Releasing the Revio first (a 4 chip system) rather than an a cheaper single chip platform also seems like mistake in hindsight. The Revio likely doesn’t (yet) have the necessary pull through to warrant a 4 chip system for many users. They’ve already reduced the price to ~$350K. This is probably what they could have sold a single chip system for at higher margin.
In addition to all this, we’ve got two potential/ongoing lawsuits…
So… what do we think of all this and the financial news?
Financials
So, revenue was flat versus Q1 2023. They have $562M in the bank. They sold less Revios than they expected to. They are going to reduce this by between $50M and $75M. I don’t quite understand how this translates to Net Loss. Net Loss for 2023 was $306.7 million. They laid off some folks last year, but then they also acquired Apton… Let’s make a guess that they can get losses down to ~$200M. That’s 2 and a bit years of runway.
Ideally I’d hope they reduce burn even more than this…
What’s Going To Go? What Should Go?
Say Goodbye to Short Read Platforms?
Much like Keith over at Omicsomics I’m a bit skeptical of their ability to execute in short reads. Or for this to solve there short term problems. Perhaps they can spin this out? Or sell it? But I’d agree with Keith that it might be time to focus on the long read platform.
If I had to choose a single genome, out of those available the Revio still seems to provide the best clinical genome at the lowest cost. I can’t really Apton-Omniome-Onso is going to give a significant over any other competitor (unless it’s cheaper, which isn’t going to solve any short term issues). So it seems like a good idea to reduce burn in the short read area… or perhaps find someone to acquire it. Perhaps made sense when cash was more freely available… but maybe not now.
Cancel Some Long Read Platforms?
That still leaves four other long read sequencing platforms.
I’d guess many of these won’t survive either. The most likely candidate for “cancellation” seems like the “SuperRevio”. It doesn’t seem like users are using the Revio to its full capacity, so no great reason to build a bigger instrument.
Sequel should survive only on chips already fabbed, unless there are long term commitments in place. After that… cancel it. So Sequel IIe goes EOL.
That leaves the future bench top instrument… what is it? We don’t know. Does it use Revio chips? If not… maybe cancel it.
Simpler Cheaper Platform
Start selling a stripped down Revio today.
From comments on the Discord and company statements PacBio seem to spend a ton of money on compute in the Revio. This should not be the case. The PacBio primary data does not require a very complex base calling model. They seem to have embedded DeepConsensus. Pull all this stuff out, but a simple base calling model in the instrument, even if it takes you down 1 or 2 Q scores. If you need a consultant to help you figure this out… I’m available. ;)
So strip out some of the FPGAs fed from those 4 chips to create a single chip system. Strip out the optics for the other chips. Leave the laser and everything else alone. Remove most of the compute..
No new engineering work. Just placement of different/fewer parts if the FPGAs really cost enough to justify it…
Lawsuits/IP Stuff
We’ve got two lawsuits, one with Personal Genomics, the other with Take2. Let’s take a look.
Personal Genomics
Personal Genomics used to be called CrackerBio. I don’t get the impression they are a large entity, I’ve exchanged emails with the CEO a few times (Hi!). The IP stuff appears to cover single molecule sensing on a chip which has a “pinhole” like structure. A number of the claims were invalidated at a IPR but several held.
I don’t know how much Personal Genomics have to fight a costly trial. Perhaps PacBio could settle this relatively cheaply and it does seem annoying to have this hanging around.
I took a quick look at the patent, and tried to figure out if the claims that haven’t been invalidated would cause PacBio significant issues…
The patent mostly says “you could make a pinhole and stick a single molecule to it on a chip” it doesn’t seem to discuss the ZMW approach. It mostly talks about sticking a biomolecule to a chip as they say “An apparatus for identifying a single biomolecule comprising...a linker site formed over the light detector, the linker site being treated to affix the biomolecule to the linker site.”. It seems clear from the specification that my biomolecule they mean nucleic acid.
PacBio don’t do this, they stick a polymerase to the chip. The specification discusses this saying that linking may be indirect via “e.g. a polymerase”. But is that how the claims are going to be read?
There’s a lot of other stuff going on in the PacBio system between the “biomolecule” and the sensor which doesn’t seem to be discussed here. There also are no practical demonstrations in the patent so perhaps PacBio can argue enablement? Who knows… always difficult to know how these things will play out.
Take2
The Take2 IP covers detection of base modifications on a system like PacBio’s. I’d not sure how this is even a thing… their IP appears to have been filed in 2020, and PacBio have IP on using the inter-pulse duration going back to 2013 at least and they published methylation detection way back in 2010.
According to GenomeWeb PacBio’s lawyers said:
“claims a statistical model for calculating an allegedly better prediction as to whether a nucleic acid base is modified," they wrote. "A statistical model is ineligible subject matter because it is essentially math, even if it solves a problem of genetic prediction.”
So What Do I Think?
I don’t do day trading and I’ve never bought shares in any sequencing company… And I don’t really have free cash to play with.
But a PacBio market cap of $375M does kind of have me tempted. So does a Oxford Nanopore market cap of $850M. Both these seem shockingly low. Both technologies have unique characteristics that make them the best/only way to address certain problems… and in both cases I can’t see the basic technological approach going away in the short to medium term.
And even from a strategic acquisition point of view… both companies seem like their worth an acquisition price >=$1B.
This of course makes a $375M market cap seem crazy low to me. Particularly given how well they’ve executed technically with the Revio. And even with all the other “baggage” mentioned above.
None of this is meant as investment advice… and very little about share prices seems rational anyway… but… for me… with my risk profile… it’s… tempting.
Maybe I’ll check round the back of the sofa again and skip coffee tomorrow.