What does the future of lab automation services look like, if there is one? Is Transcriptic a viable model? Is there enough of a market there, or are protocols too niche?
Some thoughts on Transcriptic/lab automation in general:
1. In general, what protocols are robustly and reproducibly automatable? IMO, it's really only nucleic acid work (e.g. cloning/site-directed mutagenesis/etc) - as you go up the stack in complexity (protein/cells/tissues/animals), it becomes very difficult to readily automate tasks. Subsequently, the protocols that are well-handled by Transcriptic/lab automation happen to be the ones that academic labs can do pretty well themselves and pharma doesn't care too much to do...
2. Automating cloning/plasmid work is effectively competing with DNA synthesis, which is bound to improve quickly enough to make most cloning obsolete (maybe it won't be this wave of microarray-derived DNA synthesis, ala Gen9/Cambrian Genomics, but certainly by the next generation...) - it's a dangerous area to tread.
3. This model, while interesting when pitched in a "cloud/AWS"-esque way, is not new; fundamentally, Transcriptic is a CRO, albeit with some automation (which most CROs have access to as well - if I recall correctly, Transcriptic is not developing any new instrumentation or processes, only software improvements like scheduling and UI). Given that most CROs have evaporated in the US and have largely all moved to Asia (China/Singapore/India), how well can Transcriptic compete with them (the "automation beats cheap labor" argument doesn't hold well in this case - see the current DNA synthesis market for an obvious example)?
4. Who are the customers and how big is the market? Academic labs again can largely depend on students to carry out most protocols (as a sidenote, Max did point out that he wanted to target various informatics/compbio labs that might want to empirically test their models, which is a somewhat reasonable idea). Pharma typically go with "trusted" CROs that they have long-lasting relationships with; maybe they can go with small biotechs, but a) those companies are largely in a funding crunch right now and b) Max needs to provide a much more broad spectrum of services before Transcriptic really becomes useful.
5. Also note that there is another DNA assembly startup (Teselagen, based out of JBEI's j5 software), which effectively determines which assembly process - Golden Gate vs Gibson vs Gateway vs SLIC/etc - will be most cost-effective/most likely to succeed and then sends commands down to a Biomek to complete the assembly. I think Teselagen probably stands a better chance than Transcriptic, given that their software layer is reasonably functional and that they already have decent hook-in with the "synthetic biology" community.
6. As one aside, I really do like the vision of fully automated experimentation (obviously - I played around with lab automation myself in the past...) and it'd be great to be at a point where protocols are robust enough to readily automate, so folks can start doing git-style pushing/forking/etc of protocols and overall biology can be much more "formalized". Unfortunately we're still pretty far away from that though...
3&4: I definitely see evidence for a role for small biotech-CRO hybrid operations working with pharma; there are a bunch of these operating in particular technological niche (e.g. Flagship Bio flagshipbio.com).
These players often have strong automation internally, and wield that to provide a significant operational advantage to customers. It seems to me, however, that these smaller companies market this advantage as an analytical capability that a pharma would otherwise not have, rather than an accelerant to existing research processes.
This is because the big CRO players (and the pharma customers themselves) can always wield brute force (people+robots) to beat out a smaller player at speed, even if its more costly.
Another startup taking the "analytical capability that a pharma would otherwise not have" market strategy is the Science Exchange which seems to have found a niche in doing reproducibility studies[1], which the industry has recognized to be critical for moving research from academia[2], but are costly for companies to do in-house because they are not amenable to automation.
More broadly, the research earlier and earlier in the drug development pipeline would seem to be the work that pharmas are less willing to do in-house, but might be willing to pay for if there was clear utility. Clear utility, it seems, is the hard part.
[1]: scienceexchange.com
[2]: blogs.nature.com
With that said, I, too, worry about the immediate positioning of Transcriptic for the reasons Darren has mentioned: there's no pressure or inclination for academics to use CROs for various "traditional" reasons, and targeting academia as a market is not quite pragmatic (we've explored this with Perceptus) due to fragmentation and niches. At best, academia is a way to validate technologies and beta test.
Pharma market targeting for cloning and assembly is difficult, as it's something that they already do (and big CROs do fast), and not technical niche.
In my mind, the question is how to target the transition points between "academic leads" and pharma hits; big companies will pay for a good hit (or validation of such).
It's also worth noting that successful CROs and biotechs have sprung up not in the wet-lab DNA space, but in making sense of genomic, systems biology, and molecular bio experiments. Companies have figured out how to provide unique analytical capabilities making use of this "commodity" science, and offer that as a service.
Is there a way Transcriptic can provide unique analytical services on top of its automation platform? This would be akin to Google offering specialized advertisement targeting services informed by their search engine.
