Figure 3

Trueness and precision at various dilutions. (a) Dilution table of CD1 and BL6 genomic DNA mixtures. Total DNA concentration was the same for each group. Dark gray boxes labeled NA were not analyzed because the concentration of low-abundance DNA was less than 0.5 molecules per reaction. (b) Combined Groups 1–5 concentration measurements compared to expected concentration; axes are log-scaled. 5% random error added to each point to aid in distinguishing overlapped datapoints. The red dashed line indicates the concentration at which there is only a single positive droplet per reaction (single molecule detection limit). (c) Coeffecient of variation (CoV, standard deviation divided by mean) calculated for the low-abundance allele in Groups 1–5 (n = 4 for each point: combined low-abundance CD1 and BL6 measurements, and two replicates of each). Gray dashed bar approximates theoretical CoV from subsampling error. (d) The number of false positive droplets in negative controls (n = 116). (e) Theoretical Poisson distribution of measured concentration when a sample is prepared at the LoD (6 molecules per reaction). The red dashed line and the blue dashed line indicate LoB and LoD, respectively. Black x-axis shows concentration as positive drops/reaction; brown x-axis shows concentration as copies/µl (assume 17,000 total drops/reaction). Area to the left of the LoB is β error (1.7%).