Circuit transfer functions
Each repressor gate's measured dose-response, with the points this circuit actually operates at drawn on it.
What it does
A predicted dynamic range is one number, and it hides the thing that decides whether a design works at the bench. A circuit scoring 324-fold can have every gate sitting comfortably either side of its transition, or it can be balanced on a cliff where a small change in expression collapses the separation. Both report the same number. Only the curve says which one you have.
Each plot is one gate's response function — the output promoter activity it produces for a given input promoter activity — swept across the range where that gate actually switches. On top of it sit this design's own operating points: one per row of the truth table, coloured by whether the gate is meant to be ON or OFF in that row.
What this computes
| Readout | What it is |
|---|---|
| Response curve | The gate's dose-response, evaluated from the parameters measured for that specific characterized variant. Repressors invert, so the curve falls: more input, less output. |
| Operating points | Where this circuit sits on that curve, one per input combination. The horizontal position is the input the gate receives in that row, including the contribution of both promoters where a gate reads two. The vertical position is the output it produces. |
| Logic state | Whether the gate is intended ON (green) or OFF (vermillion) in that row. Both states are shown, because a design is only a switch if the two sets separate. |
| Midpoint | The input level at which the gate is halfway between its floor and its ceiling — the centre of its transition, in relative promoter units. Each curve is swept around its own midpoint, since these differ by orders of magnitude across a library. |
How to read it
Look at where the dots fall, not just how far apart they are. Dots spread either side of the falling part of the curve mean the gate is being used as a switch. Dots bunched together at the flat top or the flat bottom mean it is saturated: it is contributing separation now, but it has no margin, and a gate with no margin is the one that fails when the real host expresses it slightly differently from the characterization strain.
Green above vermillion is backwards. These are repressors, so a gate that is logically ON has a low input and a high output. If the green dots sit at higher output than the vermillion ones, that is correct.
Units are relative promoter units throughout, and both axes are logarithmic — a dose-response spans decades, and on a linear axis everything but the transition lands on the axis line. The axis labels say so.
Occasionally one node is drawn twice. A repressor group holding several characterized variants is evaluated variant by variant for each row, and the best-performing one can differ between rows; when it does, each variant gets its own curve carrying the rows it actually produced. A point belongs to the gate that produced it, and drawing it on another gate's curve would be a picture contradicting its own numbers.