A "Referral from the Doctor" Blog Article-
The mathematical model of PCR is based on the assumption that each DNA template molecule is reliably duplicated once per cycle, assuming an excess of reagents. In practice, a variety of factors impact amplification kinetics, particularly when the copy number is low, causing deviation from the ideal model. These factors include RNA quality, residual inhibitors, operator technique, primer quality, and efficiency of reverse transcription, to name a few. In addition to those readily identified factors there is also the unavoidable 'Monte Carlo' effect.
The ‘Monte Carlo’ effect, so named because of its association with probabilities to predict outcomes in gambling, describes an inherent limitation while amplifying templates expressed at very low levels. The higher variance in the results from PCR reactions with a low starting template number (< 100 copies) contributes to this statistical phenomenon, such that more qualitative information is produced, as compared to reactions with abundant targets.
An operating theory for this abrupt increase in variation is based on the premise that a primer annealing and successfully replicating an individual template is random and governed by a probability of occurring. If a primer fails to bind then that molecule will not replicate and must await the next cycle for another opportunity. Such a misfire is easily overlooked when millions of other template molecules were successfully replicated during that same cycle, but the consequences are much more pronounced with only one or a few molecules. For this reason, the Monte Carlo effect is presumed to come into play early during thermal cycling, before the target has become enriched. The outcome can be a reduced yield of PCR product or scattered CT values between replicate qPCR reactions.
Such stochastic events must be properly accounted for while interpreting PCR results from samples with low template number or risk compromising quantification. Many more replicate reactions are required to achieve statistical confidence in the results.
Written by: Christina Ferrell, Ph.D.
- Biophysical Journal 71:101-108, 1996
- J Biomol Tech. 15(3): 155–166, 2004
- Proc. Natl. Acad. Sci. USA (Biochemistry) 92:3814-3818, 1995