Researchers at Harvard Medical School are studying the differences in human genomes to identify the genes underlying biological processes and human disease. One of the researchers, professor Steve McCarroll, is analyzing the copy number variations (CNV) of genome segments from tens to hundreds of thousands of base pairs long. This requires the ability to measure the precise copy number of these segments in thousands of individuals.
Using technologies such as real-time PCR (qPCR) and comparative genomic hybridization arrays, the McCarroll lab could measure simple deletions and duplications—changes in copy number from two to one or zero, or from two to three or four. But what they lacked was the ability to precisely and reproducibly measure copy numbers greater than four.
Bio-Rad Laboratories’ QX100 Droplet Digital PCR (ddPCR) system provides an absolute measure of target DNA and RNA molecules. It can be used to discriminate small-fold differences in copy number, enabling researchers to measure 1, 2, 3, 4, 5, 6, or more copies.
Whereas qPCR quantifies nucleic acids by comparing the number of amplification cycles and amount of PCR end-product to those of a reference sample, ddPCR enables researchers to directly quantify nucleic acids. In other words, ddPCR does not require the use of a standard curve.
Although qPCR is a viable detection strategy when mutant and wild-type sequences are mixed, its effectiveness declines when the mutated sequence is relatively rare, such as in myleoplastic syndromes (<20–25%).
Droplet Digital PCR separates samples into 20,000 droplets, and reactions are carried out individually in each. This reduces background interference for more reliable and sensitive measurement of low concentrations of nucleic acid that may not have been detectable using qPCR.
How ddPCR Works
Droplet Digital PCR takes advantage of simple microfluidic circuits and surfactant chemistries to divide a 20 µL mixture of sample and reagents into 20,000 droplets with target and background DNA randomly distributed among them. These droplets support PCR amplification of single template molecules using assay chemistries and workflows similar to those for qPCR applications (i.e., TaqMan).
After PCR amplification occurs, a reader determines which droplets contain a target and which do not. Software calculates the concentration of target DNA as copies per microliter from the fraction of positive reactions using Poisson statistics.
In digital PCR, having more partitions provides greater precision and resolution for detecting small concentration differences. Sample partitioning to levels above 10,000 allows for extremely accurate Poisson correlations and substantial enrichment effects when screening for rare events. The QX100 is the only digital PCR system that generates uniform droplets to partition a target sample. Others use microfluidic chips, which are challenging to scale to higher partition numbers per sample while maintaining low costs.
The arrival of qPCR revolutionized the field of gene expression, becoming the predominant technology for this type of research. With improved instrumentation and high-quality reagents, researchers are now able to report gene-expression levels varying at very fine amounts.
Unfortunately, resolution at levels below 50% are difficult to achieve due to the compounding of errors derived from each step in the quantification process. From a qPCR perspective, these include the dependence on a standard curve for amplification efficiency determination, standard error of this curve, technical replicate variation, normalization to reference genes (each of which carries previous errors), and comparisons between samples (e.g., normal and treated) that carry all these errors. Low expression targets tend to demonstrate greater variability between replicates due to the larger number of cycles required for amplification.
Droplet Digital PCR alleviates the compounding error effect, as readings are absolute; quantification is precise due to the digital nature of the assay and the fact that no standard curve is required. Normalization to multiple reference genes should still be performed (according to the MIQE guidelines) but here again, all values for these are standalone and not dependent on other samples or references, minimizing carried errors. Additionally, as a result of this independence, when quantifying samples on different plates, different dates, or in different labs, the use of normalizing reference samples is no longer required.