10X Scoops Scale

News broke this afternoon that microfluidics single cell genomics leader 10X Genomics has bought out split-pool single cell company Scale Biosciences for $30 million up front in cash and stock, with additional payments of undisclosed amounts if certain milestones are reached.  

Scale had been aggressively driving up the number of cells which could be analyzed in a single experiment with their QuantumScale technology (which I profiled after this year's AGBT) and ScalePlex technology to enable the large number of cells to be divided among many samples.

One appeal of such high scale is what is driving much of the demand for single cell genomics: generating data to feed into AI model training.  Disclosure/advertisement: my employer has a business division, Ginkgo Datapoints, which has a single cell service offering exactly on these lines.  More is better - more cell lines, more treatments, more timepoints, more doses, more more - you get the picture.  So Scale seemed well suited and therefore attractive to 10X.

Enabling such scaling up is the scaling down of sequencing costs.  With Ultima delivering 100 gigabase pairs for $80, which would have cost 4-5X that only a few years ago, one can dream of bigger and bigger experimental schemes.  Single cell genomics, particularly for model training, isn't as sensitive to the homopolymer issues that pop up in Ultima's reads.  

Also perhaps driving 10X's interest in Scale is the fact that Illumina last summer bought Fluent BioSciences, another company offering an instrument-free single cell solution.  10X's Chromium Controller instrument offers very high quality partitioning of cells into droplets and is clearly the industry leading technology for single cell analysis, but a physical instrument is a bottleneck.  Scale's technology relies solely on liquid handling, and Fluent's nearly so (there's a vortexer of some sort in the Fluent scheme).  

There's often been speculation that 10X would respond to Illumina's encroachment on their single cell (and now spatial) turf by making an acquisition in Illumina's sequencing space, such as Ultima or Element.  While the Scale acquisition doesn't preclude such a move in the future, it does suggest that 10X's management doesn't see that as quite as attractive a possibility - though perhaps they still have enough bank account to not be forced to sell for such an attractive price.

10X also has a reputation for an aggressive legal strategy with their patents, and now they have Scale's patents.  That can't make Parse Biosciences very happy.  Parse and Scale have very similar split-pool workflows and much of the technologies came out of adjacent labs at University of Washington.  And for whatever reason of just random chance or perverse humor, at AGBT this year Scale & Parse were in adjacent suites at the JW Marriott! 10X and Parse have engaged in a series of patent litigations that have seen each side win some and lose others; I haven't kept a scorecard and I suspect only someone deeply immersed in these patents would truly understand the significance of each reported outcome.  

An interesting note is that $30M sale number.  According to Google's Gemini AI, Scale had raised $32.5M in venture funding.  So the investors are getting to save a little face by almost getting their money back.  That Scale was valued so low says a lot about the dismal state of biomedical science environment right now, with biopharma and industrial biotech companies slashing jobs right-and-left and the current US Administration gunning for maximum chaos in the academic funding sphere.  AI-oriented startups seem to be almost the only thing venture capitalists fund these days, but that didn't keep Scale going.  Selling now suggests that Scale, which was private, was probably running low on cash and would need to raise funds soon - an absolutely terrifying task in the current environment.

Single cell is also interesting because there are still new approaches coming out of academia and other startups (such as CS Genetics) trying to break into the space.  Single cell also competes with spatial in some ways, particularly with the STAMP method of packing cells on a slide and then probing spatial-style - generating single cell information on large number of cells.  There's also a whole series of companies, such as Atrandi, Cellanome, and others, which are trying to enable generating single cell profiles on much smaller numbers of cells but with a great deal more imaging on the cells to capture rich phenotypic and perturbational information.

Single cell technology will continue to evolve, and with it the landscape of tools companies in this space (and I've admittedly left MGI out of the discussion as I believe their inexpensive microfluidic tools are not sold in the US for IP reasons).  Whether the current AI gold rush will continue is a major question in this field, or whether demand will return to be entirely driven by basic research.  Of course, any company in the tools space dreams of getting their tool into clinical diagnostics, but that's a serious challenge for single cell - first in finding a companion diagnostic that really requires single cell and then the very real possibility that any such assay would be converted from exotic single cell to a conventional histology assay - which are, after all, single cell spatial readouts!