Direct elicitation of template concentration from quantification cycle distributions in digital PCR

by Mr Aymeric Fouquier d'Herouel (KTH Computational Biology)

Tuesday 14 May 2013, 13:30 → 14:30 Europe/Stockholm

122:026

Digital polymerase chain reaction (dPCR) offers a robust means to quantify absolute numbers of template molecules by partitioning a sample over multiple replicate reactions. The number of replicates is chosen such that each reaction contains zero or one template. The number of successful amplifications ideally corresponds to the number of templates in the partitioned sample. When partitioning a sample, it is not possible to guarantee that each volume contains at most one template; the contents of individual reaction volumes follow Poisson statistics and the presence of multiple templates per reaction leads to undercounting effects. In practice, this is accounted for by applying a simple statistical correction to measurement results.
Although dPCR results are essentially endpoint measurements, complete real-time amplification curves are usually recorded by instruments. The notion of quantification cycles as in qRT-PCR thus also holds in dPCR. I will discuss a method that exploits the availability of such data by comparing the measured distributions of quantification cycles with those expected from an analytical model of PCR, yielding the absolute template concentration in the initial sample. The method circumvents the need for post-hoc corrections while relying on a constrained optimization procedure and compares well with the most common statistical correction in preliminary results.