Stochastics of DNA Quantification

TL;DR

This paper introduces a stochastic mathematical model for DNA quantification via PCR, accounting for amplification noise, and provides new formulas for detection limits, improving accuracy over traditional methods.

Contribution

It develops a parsimonious stochastic model of PCR amplification noise and derives formulas for detection and quantification limits, enhancing the accuracy of DNA quantification analysis.

Findings

Model predicts <4% positive detection for single DNA molecule

Provides formulas for detection and quantification limits

Shows standard methods are less accurate than the new model

Abstract

A common approach to quantifying DNA involves repeated cycles of DNA amplification. This approach, employed by the polymerase chain reaction (PCR), produces outputs that are corrupted by amplification noise, making it challenging to accurately back-calculate the amount of input DNA. Standard mathematical solutions to this back-calculation problem do not take adequate account of such noise and are error-prone. Here, we develop a parsimonious mathematical model of the stochastic mapping of input DNA onto experimental outputs that accounts, in a natural way, for amplification noise. We use the model to derive the probability density of the quantification cycle, a frequently reported experimental output, which can be fit to data to estimate input DNA. Strikingly, the model predicts that a sample with only one input DNA molecule has a $<$4% chance of testing positive, which is $>$25-fold lower than assumed by a standard method of interpreting PCR data. We provide formulae for calculating both the limit of detection and the limit of quantification, two important operating characteristics of DNA quantification methods that are frequently assessed by using ad-hoc mathematical techniques. Our results provide a mathematical foundation for the rigorous analysis of DNA quantification.