There is a core concept in the land of silicon that is especially interesting when applied to biology: the idea of a generic processor. While you can have chips that are good specifically for a single task, you also have processors which execute a small number of different commands that can be chained together in arbitrary ways. These generic processors allowed for programmable computers to become widespread and commonplace.
Right now, biology labs are very much like chips. They are configured to do certain protocols, and do those protocols well. While individual labs and individual lab protocols might be able to be chained together in arbitrary ways, those instructions can’t be easily translated or replicated to other labs, creating a sort of technological island that companies or academic PIs love to claim ownership of.
Specifically, the physical manipulation of tubes, liquids, and equipment required in biology protocols (and the tacit knowledge necessary for successful execution) are not easily replicated across physical spaces. This disaster, while a known and acknowledged problem by scientific biologists, at its core is not widely acknowledged as a problem for bioengineers. The most unfortunate part of this situation is that it is up to the engineers to fix it, and right now they aren’t.
Reproducibility isn’t just a scientific excerise: reproducibility allows for greater collaboration and through increased collaboration (and thus demand) drops the price of participation. Biological reproducibility is not just about finding scientific truth, it is also about creating a more equitable future where more people can participate in engineering life.
We need better interfaces for reproducing experiments
There are two simple engineering feats that need to be accomplished for biological protocols to achieve engineered reproducibility:
- A standard, cross-platform, and reliable interface for doing a protocol
- Objective metrics and standards for the successful completion of a protocol
I’m by no means perscriptive with these two feats, but I am taking from the lived experience of many software engineers. Software engineers are freed from many stupid problems by the availability of reliable interfaces and objective metrics, and, importantly, are happy to loudly complain when interfaces change or metrics fail. From bioengineers, I more often hear some kind of abdication of responsibility for failing to standardize (“it’s too hard”, “biology is too weird”, “I don’t have time for that”) rather than an honest attempt. The first party might be right, biology might be too hard to standardize for, or it might be that we just aren’t complaining loudly enough.
We need labs to operate like computers
We need a set of different interfaces, constituting a lab, with biological foundries developed beneath those interfaces to implement that lab. Engineers and scientists could creatively instruct the foundries to physically do their experiments without the need for the tacit knowledge of how a protocol works or any other physical lab knowledge. Objective metrics mean that these interfaces would stay the same for scientists and engineers even if they moved across the world or started their own lab.
You don’t need to know how to program in assembly to use a computer, and someday, you won’t have to know how to use a pipette to do biology experiments. That is dream of The Computer (biology) Lab. It is a lab whose instructions, not people, could be sent across the world and perfectly reimplemented. A lab whose prowess isn’t one of locality but of creativity. It exists in a world where you don’t need to have your own lab or know what a microcentrifuge tube is to build beautiful things with biology.
We need leadership to make things better
There is evidence that the world isn’t ready for the kind of innovation that will make Computer Labs (see Transcriptic), but I think that the largest obstacle is a little different. We are stuck in a local maximum of productivity, and it is going to take real courage to get out. It’s going to take a fearless advocate, a person with a vision of what the future could be, to convince enough people to embark across that dangerous valley to bring us to a new mountain of productivity. It will take many people to fully realize this dream, but it starts with single individuals whose imagination and inspiration will shape our future.
[PS: Thanks to Dhasharath Shrivathsa at Radix Labs for the inspiration of thinking about labs as computers. I’ve changed the idea a little in my head, but he is the one brought this idea to light for me]