EC50, IC50 and Ki all describe how much of a compound it takes to do something, and they are not interchangeable — two depend on the assay and one does not. An EC50 is the concentration producing half of a compound's maximal response; an IC50, half of its maximal inhibition; a Ki is a binding constant, a property of the molecule and its receptor alone. Comparing one paper's EC50 with another's Ki compares an experiment with a molecule.
The three, defined
| EC50 | IC50 | Ki | |
|---|---|---|---|
| What it is | Concentration for half-maximal effect | Concentration for half-maximal inhibition | Equilibrium dissociation constant for binding |
| Measured by | A functional assay: cAMP, calcium, a response | A functional or competition assay | Binding, usually competition against a labelled ligand |
| Depends on the assay? | Yes, strongly | Yes, strongly | No, in principle |
| Depends on receptor density? | Yes | Yes | No |
| Comparable across papers? | Rarely | Rarely | Usually, if the receptor and ligand match |
Why does an EC50 depend on the assay?
Because a functional response is amplified between the receptor and the readout, and the amplification varies. A cell with many receptors and a strong downstream cascade reaches half its maximal response with only a small fraction of receptors occupied, so the EC50 sits well below the binding affinity. A cell with few receptors needs most of them occupied and the EC50 approaches the affinity. Same compound, same receptor, different EC50 — because the cell line, the receptor expression level, and the readout chosen all sit between the binding event and the number. This is the reason incretin potencies do not travel between papers, and it applies to every functional assay.
What makes Ki different?
It describes the binding equilibrium itself: the concentration at which half the receptors are occupied, with nothing downstream in the way. It is measured by competition — the test compound displaces a labelled ligand of known affinity, and the Ki is calculated from how much displacement occurs at each concentration. Done properly, it is a property of the compound and the receptor, and two laboratories measuring it on the same receptor should agree. What it does not tell you is whether the compound activates the receptor once bound: an antagonist and an agonist can have identical Ki values and opposite effects.
Reading potency claims
Three habits prevent most misreadings.
Identify which number it is. "Potency" in a paper may mean any of the three, and the abstract often does not say. The methods section does.
Identify the system. For an EC50 or IC50, the cell line, the receptor expression, the readout, the serum or albumin content, and the exposure time. A number without those is a number without units in any meaningful sense.
Compare within a paper, not between them. Two compounds ranked in the same assay under the same conditions are genuinely ranked. The same two compounds each measured in its own paper are not, and the difference between their reported EC50 values may be entirely assay.
Why do EC50 values for peptides vary so much between papers?
For the general reasons above, plus three that belong to peptides. The nominal concentration is often wrong, because the stock was calculated from label mass rather than net peptide content. The actual concentration is lower still, because of adsorption at the dilute end of the series and degradation in the medium. And for an acylated peptide, albumin in the medium sequesters most of the compound, so the free concentration bears little relation to the nominal one. Each of these shifts the apparent EC50 rightward by an unmeasured amount, and different laboratories are wrong by different amounts.
What is the difference between IC50 and Ki in a binding assay?
The IC50 in a competition binding assay is the concentration that displaces half the labelled ligand, and it depends on how much labelled ligand was present — more competitor is needed to displace more. The Cheng–Prusoff correction converts that IC50 to a Ki by accounting for the labelled ligand's concentration and affinity. A paper reporting a binding IC50 without the correction has reported a number specific to its own assay conditions; one reporting Ki has, if the correction was done properly, reported a property of the compound. The two can differ several-fold, and the label matters.
What is a Hill slope, and why does it appear next to an EC50?
The steepness of the curve, and a diagnostic of what kind of curve it is. A slope of 1 is the simple case: one compound, one binding site, no cooperativity. A slope well above 1 suggests cooperativity, aggregation of the compound at high concentration, or a threshold effect. A slope well below 1 suggests more than one site, more than one population of receptors, or a compound that is partly depleted at low concentrations — which for a peptide usually means adsorption. An EC50 from a curve with an odd slope is an EC50 from a curve that may be describing something other than simple binding, and the design of the series is where that gets diagnosed.
Which figure should a methods section report?
Whichever was measured, labelled as such, with the system it was measured in. An EC50 with its cell line, readout, serum content, exposure time and the number of independent curves behind it is a complete result. A Ki with its receptor source, labelled ligand and whether the Cheng–Prusoff correction was applied is a complete result. "Potency of 3 nM" is neither, and a reader cannot tell which of the three it is or what it can be compared to. The word potency is where the information goes missing; the specific term is where it stays.
All products referenced here are supplied for laboratory and research use only. They are not drugs, foods, supplements or cosmetics, and are not for human or veterinary use.
