Biotech automation sucks
Building instrumentation around research (not vice versa) - this is the way
Biotech R&D is still a highly manual process - while individual instruments have become much higher throughput, the workflow of most research still requires people to move vials here and there, to put plates in incubators, and do other repetitive tasks. Why is so little of bio research automated?
Biology is an observational discipline, not yet an engineering one. The reason our “DNA code is software” analogies break down pretty quick, is that we are working with insanely complex systems that evolved over billions of years, not systems that are manmade. Once we are able to recapitulate a cell from scratch (see: the Build a Cell movement), then biology will begin to be an engineering discipline… but we are not there yet.
Because most of biology focuses on deep study of existing systems, especially in the R phase of R&D, most biologists are trained in observation. As undergrads, grad students, and post-docs we use tools to probe complexity that developed over eons. And when it comes time to move to the development stage, to start companies, to “translate” biology, we still use these observational tools - it is this instrumentation that underpins bio labs from high schools to big pharma: microscopes, centrifuges, PCR machines, and chromatography. These instruments were not designed to be automated, they were designed for manual, observational research.
Hacking automation on top of this system is challenging and incremental. It seems the best we can do is robot arms on rails moving plates and tubes around. There is no holistic solution for automating biotech once research is locked into using classic instrumentation.
This is why so many companies have struggled. Not only are the instruments not conducive to automation, the format of research and development is not conducive to automation, because it has been shaped and driven by the use of this instrumentation.
So what can we do?
Design R&D from first principles, and build instrumentation around the work that needs to get done - designing with the end (automation) in mind from step 1 of your R&D plan
Create the spaces where this work can happen - incubators, non-academic labs, venture studios, community bio spaces, or in startups. Get investors willing to support whole cloth R&D systems redesign - this requires substantially more up front investment.
Redo protocols by cutting steps and complexity - no references to “how it’s always been done”
Foster talent and innovation outside of traditional institutions - the DIY bio community, for example, has the freedom to reinvent research, without decades of institutional knowledge weighing them down, and are adept at building cheap instrumentation.
With an increase in R&D focused instrumentation design, and not instrumentation driven R&D, we can truly start to implement automation in biotech. We need to drive costs down to (finally) unlock exponential growth in biotech products.