User manual

Absolute Quantification and the Statistics of Droplet Digital
™
PCR
30 | Droplet Digital
™
PCR Applications Guide
Fig. 3.2. Sample concentrations are plotted as copies/µl from the sample.
0 256 256 64 64 16 16 4 4 1 1
Sample
Concentration, copies/µl
10,000
1,000
100
10
1
0.10
0.01
1,390 1,400 1,400 1,400 1,360 1,370 1,380 1,390 1,370 1,360
260 259
64.2 64.8
15.5 15
4.63 4.68
1.46 1.53
Statistics of ddPCR
In ddPCR, DNA molecules and PCR reagents are partitioned into droplets. Some droplets
will contain one or more copies of the target DNA molecule (DNA fragment that contains
primer and probe binding sites for a given assay) and some will not contain any copies of
the target. Droplets are subjected to PCR amplification and the targeted PCR product, with
an associated fluorescence signal, is generated in droplets that initially contained the target
DNA. The concentration of the target species in the original sample is determined from the
fraction of droplets that have a fluorescence signal.
In ddPCR, target molecules are distributed randomly into droplets. This assumption is at the
heart of all digital PCR approaches. Random partitioning means a given target molecule is
equally likely to end up in any of the 20,000 droplets. Random partitioning also means target
molecules move independently of each other and do not interact with each other.
The droplet generator partitions samples into approximately 20,000 droplets of identical
volume (about 1 nl). Some droplets are lost in transfer steps and others are eliminated by
the stringent metrics applied by QuantaSoft software as the droplets pass through the
droplet reader, resulting ultimately in data from 12,000–16,000 droplets being used in
subsequent concentration calculations. Reading only a subset of the total droplets has
no impact on the concentration measurement, because the concentration is calculated
based on the fraction of droplets that is empty (that is, the fraction that does not contain
any target DNA).
As an example, suppose a sample contains 1,000 copies/µl of a target molecule, and the
sample is loaded into two ddPCR wells. In one well, only 10,000 droplets are read by the
software, and in the other, 15,000 droplets are read. In both cases, the fraction of empty
droplets will be the same, and thus the concentration in copies/µl will be the same in both
wells (within the limits of the instrument measurement error).
At the extreme ends of the concentration range (for example, fewer than ten copies of target
in a well or more than 120,000 copies of target in a well), fewer droplets in a well will lead to
slightly larger error bars.
Note: For simplicity, we will use 20,000 partitions in the examples in this chapter.