Synthetic biology

Synthetic biology is an engineering premise applied to living matter: that biological function can be decomposed into parts with defined behaviour, that parts compose into devices and devices into systems, and that a designer can therefore work at one level of abstraction without holding the level below in mind. That premise is what separates it from genetic engineering, which modifies one organism for one purpose. It is also the premise that biology keeps refusing, and the interesting content of the field is in the refusals.

What a part is supposed to be

The canonical parts are transcriptional: a promoter with a defined strength, a ribosome binding site setting translation rate, a coding sequence, a terminator. Assembled in combination they give the switches, oscillators and logic gates that made the field’s early reputation. Repression and activation are the wiring; the transfer function of a repressor — how output falls as input rises, and how sharply — is the component datasheet. Metabolic engineering uses the same vocabulary for enzymes in a pathway, with flux where a circuit has signal.

Four reasons composition fails

Context dependence. A part’s measured behaviour is not intrinsic. The mRNA sequence around a ribosome binding site folds into structures that occlude it, so translation rate changes when the upstream sequence changes; promoter output depends on supercoiling and on what is transcribed next to it. Insulating sequences and ribozyme-based cleavage of the message exist precisely to restore a part to something like its datasheet.

Burden. A cell has a finite supply of RNA polymerase, ribosomes, charged tRNA, ATP and reducing equivalents. A construct expressing a heterologous protein at high level draws on that pool, so two circuits sharing a cell are coupled through resources even with no designed connection between them: inducing one lowers the output of the other. Burden is measurable as growth-rate reduction, and growth rate feeds back on everything, including dilution of the circuit’s own products.

Noise. Transcription is bursty and molecule numbers are small, so identical cells in identical conditions give a distribution of expression levels rather than a value. A design that works on the population average may be bimodal at the single-cell level.

Evolution. This is the failure mode with no analogue in other engineering. A burdensome construct imposes a selective disadvantage, so any mutant that inactivates it grows faster and takes over the culture. Repeated sequences invite recombination; transposons land in strong promoters. Genetic stability over the number of generations a fermentation actually requires is a design specification, not an afterthought.

Where the field goes when the cell is the problem

Two responses follow directly. One is to simplify the host until its behaviour is predictable — the reduced-genome and synthetic-cell work covered on its own page. The other is to leave the cell out: cell-free extracts run transcription and translation without growth, burden or evolution, at the cost of finite lifetime and no self-replication.

Containment is the field’s other standing obligation, and it is treated separately in the pages on genetic switches and on ecological engineering. This page deliberately does not restate the editing chemistries; those are tools synthetic biology uses, not what it is.