Evaluating digital PCR for the quantification of human nuclear DNA: determining target strandedness

Digital polymerase chain reaction (dPCR) methodology has been asserted to be a "potentially primary" analytical approach for assigning DNA concentration. The essence of dPCR measurements is the independent dispersal of fragments into multiple reaction partitions, amplifying fragments containing a target nucleotide sequence until the signal from all partitions containing at least one such fragment rises above threshold, and then determining the proportion of partitions with an above-threshold signal. Should originally double-stranded DNA (dsDNA) fragments be converted into two single strands (ssDNA) prior to dispersal, the dPCR measurements could be biased high by as much as a factor of two. Realizing dPCR's metrological potential therefore requires analytical methods for determining the proportion of ssDNA in nominally dsDNA samples. To meet this need, we have investigated several approaches to this determination: A₂₆₀ ratio, dPCR ratio, cdPCR staircase, and ddPCR enzyme. In our hands, only the endonuclease-based approach provides adequately accurate estimates for relatively small ssDNA proportions. We present four (enzyme, assay) pairs that provide self-consistent results for human nuclear DNA extracts. However, the proportion of ssDNA differs by as much as 50% between assays, apparently related to the guanine-cytosine (GC) content of the fragment near the assay's target sequence. While materials extracted by us have no more than 6% ssDNA content even after long storage, a commercially obtained PCR assay calibrant contains ≈18% ssDNA. Graphical abstract.