At the end of the year, a paper was published in Nature Chemistry detailing work done by Eterna in collaboration with the Das, Greenleaf, and Khatri labs that has some exciting implications on global health, low-energy computing, and biosensing at large. Let's talk about why this is exciting!
Sensors are all over the place in biology - living organisms are carefully tuned systems in a constant state of internal regulation, managing levels of necessary molecules, responding to environmental stimuli and defending against pathogens, and even turning genes on and off. Understanding these systems is crucial to our understanding of biology - and utilizing it to improve our health. But we can go yet one step further, using these same mechanisms to design our own custom biomolecular computers.
Tuberculosis is one of the world's most significant public health challenges, with one of the biggest bottlenecks being in diagnostics, rather than treatment - in particular, determining the presence of active (rather than latent) TB. In 2016, a three-gene signature in the blood was discovered for active pulmonary tuberculosis, leading to the creation of the Eterna OpenTB initiative. The goal? Create a fast, low-cost, easily-deployable diagnostic for TB. How? Through designing a single RNA molecule that could act as a calculator.
Such a molecule would, in the presence of either more of two of the genes (A*B) or more of the other gene (C^2) bind to another molecule that would either cause it to light up (visible under a microscope) or express a pigment (like in an antigen test). Players delivered highly effective designs - and strategies used to develop an automated algorithm for RNA sensor design, Nucleologic. The insights generated by Eterna participants were crucial - their exploratory process and unique perspective is what created the heuristics needed for the success of this automated mechanism which can further accelerate design and deployment of technologies using this approach. As a fun anecdote - the initial simpler puzzles stumped researchers when they first tried it, but players were able to solve it within days.
The simplicity of these designs is their strength - no expensive machinery to run the tests, no advanced training required to interpret results, and limited interacting parts supporting simple, reliable deployment. Similar gene signatures are being found for other diseases as well, including sepsis, cancers, and malaria, giving this approach broad potential. The paper also describes how this could lead to impacts beyond diagnostics, like directed drug delivery and safe precision gene editing (see also Eterna's work in OpenCRISPR). You could even imagine how this could be used for something like environmental monitoring (think water quality or molecule presence). The relative simplicity and efficiency also leads to a promising direction for computing - biological computers that not only operate on a scale that could solve the Feynman grant prize of developing a nano-scale computing device, but also revolutionize low-energy computing.
This is just the start of this new promising paradigm for biologic computation, with challenges ahead for designing further, more complex sensors and calculators as well as translation into deployable technology. We already have plans in the works for the next steps in this work, and are excited to take on future collaborations that will allow for these micro-scale computers to have macro-scale impact.