There are a number of ways to multiplex, but one of the most common relies on solution-based arrays of microscopic beads measuring several microns in diameter.
Like planar microarrays, these arrays are addressable - that is, each location within the array is known. But in this case, the "array" (1) is really a set of coded microspheres, each of which has an identifying color and associated bioreceptor (e.g., antibody, oligonucleotide, receptor, or enzyme). One color class might be reserved for IL-2, say, while another is reserved for IFN-gamma. Or they may represent different SNPs.
The array is mixed and incubated with a biological sample (2), after which a detection reagent (a dye-conjugated secondary antibody, for instance) is applied (3). The beads then pass single-file through a flow cytometer, which reads the reaction using two lasers (4).
The first laser induces the bead to fluoresce, thereby identifying the reaction, while the second laser quantifies the analyte itself. The results for each bead class are then aggregated, yielding a report of average fluorescence intensity for each bead type - numbers that can be converted to absolute concentrations using a standard curve.
The degree of multiplexing (that is, the number of different compounds that can be measured at once) in bead-based multiplexing is limited only by the number of colors the cytometer can resolve. Luminex's xMAP system can resolve up to 100 different bead types. Quantum Dot, prior to its acquisition by Invitrogen, devised a variation on this approach in which 200-member microsphere arrays (made using combinations of differently colored quantum dots) lying on the bottom of flat, transparent microtiter plates, were quantified directly via image processing, rather than via flow cytometry.