Single molecule array · digital immunoassay

How Simoa counts single protein molecules

A sandwich immunoassay, like ELISA, read in femtoliter microwells. At low occupancy, most beads carry zero or one labeled complex. Confining the enzymatic product makes it possible to count positive wells and detect very low concentrations. The limit of detection depends on the analyte, background and assay validation.

01 · The core idea

Same enzyme, a billion times less volume

~100 µL well
Conventional ELISA. At low concentration only a handful of enzymes are bound. Their fluorescent or colored product disperses through roughly 100 µL and never rises clearly above background.
oil seal ~40 fL wells, one bead each 1 enzyme → 1 bright well
Simoa. The same single enzyme, sealed in a well of about 40 femtoliters, increases local product concentration. Under suitable assay conditions, the signal from one enzyme can be distinguished from background.
02 · The workflow

From plasma to a fluorescence image

Step through the assay. The first four steps happen in suspension on paramagnetic beads (about 2.7 µm, roughly 500,000 per test); the last two happen on the array disc.

paramagnetic bead
bead antibody target protein SβG enzyme resorufin

03 · The readout

Counting wells, and why Poisson matters

Because there are far more beads than target molecules at low concentration, each bead carries either zero or one enzyme most of the time. The number of enzymes per bead follows a Poisson distribution with mean λ. The instrument measures the fraction of bead-containing wells that are fluorescent, fon, and converts it to average enzymes per bead (AEB).

digital: AEB = −ln(1 − fon) (used while fon ≲ 0.7)
analog: AEB = fon × Īon / Īsingle (used at higher load, where counting saturates)

Move the concentration slider and watch the array, the occupancy histogram, and where you sit on the calibration curve. A small nonspecific background is included, which sets the floor that defines the limit of detection.

no bead loaded bead, off bead, on (brighter = more enzymes)
λ true
f on
AEB measured

Enzymes per bead (Poisson, share of beads)

0123456+

Model assumptions, chosen to be in a realistic range: 100 µL effective sample, 500,000 beads, 50% capture efficiency, about 60% of wells receiving a bead. Real assays differ by analyte, antibody pair and dilution, so read the axis as illustrative, not as a specification.

04 · The physics of confinement

Shrink the volume, keep the enzyme

One streptavidin β-galactosidase molecule turning over substrate for 30 seconds makes on the order of 104 resorufin molecules (assuming ~300 turnovers per second). What that means as a concentration depends only on the volume it is trapped in.

1 aM1 pM1 nM1 µM
Resorufin after 30 s, one enzyme

05 · In practice

What to keep in mind when reading Simoa data

Calibration is per analyte

AEB is converted to concentration with a calibrator curve (typically a 4PL fit). Lot changes in beads, detector antibody or calibrator can shift absolute values, so bridging and internal controls matter in longitudinal cohorts.

Matrix still matters

Single-molecule counting removes the volume problem, not the biology. Heterophilic antibodies, hemolysis, freeze–thaw history and sample dilution all act before the bead reaches the well.

Two-step or three-step

Some kits add capture and detector antibodies together with sample; others separate the steps. Format affects hook effect risk, incubation time and sensitivity, and is defined in each kit's protocol.