Two labs can study the very same receptor, run experiments that look nearly identical at the bench, and still walk away with answers to completely different questions. The reason is design. A saturation binding assay and a competition binding assay share the same glassware, the same law-of-mass-action foundation, and often the same labeled molecule — yet each is built to extract a different number. These are in-vitro laboratory characterization methods, used for research use only, and knowing which design produced a given value tells you exactly what that value can and cannot claim.
Read enough research literature on any compound and four abbreviations keep surfacing: Kd, Bmax, IC50, and Ki. Two come most naturally from a saturation design, two from a competition design. This guide walks through what each assay measures, how they differ mechanically, the equation that bridges them, the constraints they share, and how researchers decide which one to run.
What each design is built to measure
Here is the plain-English split. A saturation assay answers: how tightly does this labeled molecule bind, and how many binding sites are there? A competition assay answers: how well does some other, unlabeled molecule compete for that same site? Same target, two different curiosities.
Four parameters carry the answers. Kd is the equilibrium dissociation constant of the labeled ligand — the concentration required to occupy half the available sites, and the standard measure of its affinity. Bmax is the maximum number of binding sites in the preparation, usually reported as something like pmol per mg of protein. IC50 is the concentration of an unlabeled competitor that knocks down half of the labeled binding under one set of conditions. And Ki is that competitor's own equilibrium dissociation constant — the condition-independent affinity you actually want to report.
Both designs rest on the same two foundations: the law of mass action, and a clean separation of specific binding (real receptor interaction) from nonspecific binding (label stuck to filters, tubes, or membrane debris). For the underlying mechanics of how a labeled probe reports affinity in the first place, our primer on radioligand binding assays covers the Kd and Ki fundamentals. The definitions above follow the NIH Assay Guidance Manual.
The saturation design: vary the label
In a saturation experiment you raise the labeled ligand concentration step by step and watch specific binding climb toward a ceiling. At low concentrations, plenty of empty sites mop up the label. As concentration rises, sites fill and the curve bends over into a plateau. That plateau is Bmax, and the concentration at half-plateau is Kd.
Practically, researchers test a range from roughly one-tenth of the expected Kd up to more than ten times it, so the curve is well defined on both sides of the midpoint. Plotting bound against free label produces a hyperbola, which nonlinear regression fits to pull out Kd and Bmax simultaneously from a single experiment — a real efficiency of the design, per the Assay Guidance Manual.
One guardrail matters throughout: no more than about 10% of the total added label should end up bound at any concentration tested. Cross that line and the free concentration you assumed no longer matches reality — the label is depleted — and both Kd and Bmax drift. There is also a shortcut version called homologous competition, in which an unlabeled copy of the same molecule serves as the competitor; that special case can still recover both Kd and Bmax, as one radioligand study illustrates.
The competition design: fix the label, vary the challenger
The competition design flips what moves. Here the labeled ligand sits at a single fixed concentration — at or below its Kd — while you steadily raise an unlabeled challenger until it displaces the label from the sites. Read out the remaining specific binding at each challenger concentration and you get a descending sigmoidal curve whose midpoint is the IC50.