Circuit design flow
This is the manual. Every tool it describes runs in your browser, free, no account.
Open the workbenchEvery stage between a Boolean specification and a DNA sequence, each showing what it actually produced for this design.
What it does
Designing a genetic circuit is not one calculation. A specification becomes a network of logic gates, those gates are matched to real repressors that have been measured in a real strain, the cassettes are ordered to satisfy the constraints the part library declares about them, and only then is there a sequence. Each of those steps can be the one that limits the result, and a single dynamic-range number at the end says nothing about which.
This view draws the chain and puts each stage's own output on it. It is also where the specification's assumptions live: the strain the parts were characterized in, the growth medium and temperature, and how many characterized gates, sensors and reporters the library actually contains. A design is a claim about those conditions, and it travels with them.
The specification is given here as a Boolean function or a truth table, and the gate network is synthesized from it directly. Both notations describe the same thing — which output each combination of inputs should produce.
What this computes
| Readout | What it is |
|---|---|
| Specification | The number of inputs and the number of rows in the truth table being realised. Every row is a condition the finished circuit has to satisfy. |
| Logic synthesis | The gate network built from that specification, counted by type. Repression is naturally inverting, so these libraries are built from NOR and NOT gates: an AND costs three gates and an exclusive-OR six. The function is minimised first, because every gate consumes a distinct repressor and gate count is the real limit on what a library can express. |
| Gate assignment | How many distinct ways of mapping repressor groups onto the network were evaluated, and how many tied for the best score. A large number of ties means the choice was not forced, and a different assignment would perform about as well. |
| Placement | How many of the constraints the library declares about its own parts this arrangement satisfies — the rules governing which cassettes may sit next to each other, in what orientation, and against which genomic landing pads. When they cannot all hold at once, the view says so rather than offering an order that quietly breaks one. |
| DNA | The number of devices in the finished construct and its total length in base pairs. |
| Gate technology | Which family of repressors the library characterizes, in the library's own words. |
| Host, genome, media, temperature | The conditions the parts were measured in, from the library's own header. A predicted level is a prediction for these conditions. |
| Characterized parts | How many gates the library holds, in how many repressor groups, with how many sensors and reporters. A group with several members gives the assignment search genuine choice; a library of single-member groups does not. |
How to read it
Read it as a funnel, and ask which stage is the narrow one. A circuit that performs poorly because synthesis needed six gates from a library holding eight is a different problem from one that performs poorly because every assignment scored about the same. The first is answered by restating the logic, the second by characterizing better parts.
A large tie count is a result, not a detail. It means the library cannot distinguish the options, so the winning assignment is best by a margin too small to rely on.
Placement is a constraint satisfaction, not a layout preference. The rules come from the part library itself and describe physical requirements: roadblocking between adjacent promoters, orientation, and which landing pad a cassette may integrate into. An arrangement satisfying fewer of them is not a tidier drawing, it is a construct more likely to misbehave.
The conditions beside the flow are part of the prediction. Every level this tool reports is in relative promoter units measured in the strain, medium and temperature shown. Build the design in a different host and the numbers are an extrapolation, not a measurement.