A lot of the inspiration, terminology, and even scientists in synthetic biology come from electrical engineering and computer science. Hence, we talk about “gene circuits” and “cellular computing”. However, this connection to digital computing hasn’t been much more than a metaphor, at least not until Cello walked on stage.
Synthetic biology has been about as much art as science, with most gene circuits requiring clever hand designing and rounds of tinkering to get them to work. Now, in a major new opus, an MIT ensemble headlined by Alee Nielson and conducted by Chris Voigt has premiered Cello (Cellular logic), an automated tool that designs genetic circuits to sense some number of environmental inputs and react with desired genetic outputs (i.e., gene expression). For now, the circuits are based on linking transcriptional repressors, whose NOT and NOR logic gates can be combined into any other Boolean gate.
In principle, a Cello user just has to specify available genetic parts (e.g., sensors and transcription factors), design constraints (e.g., what organism the circuit will be used in), and the desired logic truth table (mapping inputs to outputs). After stochastically searching for a combination of parts that should work, Cello composes a DNA sequence that can be synthesized and put into cells for testing.
Cello: An Ode To Automated Synthetic Gene Circuit Design
The work started with a prelude where the ensemble tested each of 16 different repressors to measure their input/output curves. Next, they went into a slower, minor-keyed adagio, wherein the first round of designed circuits convincingly failed. In an iterative rondo of troubleshooting, they came up with four design principles to cut down on failures:
- Insulate all genes with strong terminators
- Use ribozymes (self-cleaving RNA sequences) to make standard 5’ UTRs in the mRNAs, regardless of promoter
- Place a standard 15 bps of DNA upstream of each promoter to insulate it from upstream sequences
- Use different DNA sequences for terminators and ribozymes to avoid losing circuit parts to homologous recombination
After that elaboration, the ensemble returned to their theme by testing 52 more complex circuits as composed by Cello. 37 of these worked completely as expected, and 92% of the input-output states across all circuits were correct.
Of course, Cello doesn’t work 100% yet, and it can only design circuits for static inputs and outputs. However, as the first real design automation in synthetic biology, it’s a huge step forward.
For more, you can get your own circuits in tune at Science, 2016