To the editor:

In his commentary, W.P. Halford (Nat. Biotechnol. 17, 835, 1999) proposes that most quantitative PCR applications require neither the use of internal standards (competitive PCR), nor a quantification of the DNA products during the log-phase (TaqMan). His method relies on the occurrence of the primer-dimer artifacts in PCR reactions that compete with template DNA. The ratio of input DNA/primer-dimers will determine the amount of template DNA amplified at the plateau phase, thus providing DNA quantification if primer-dimers were constant.

Halford's commentary could be misleading to the nonspecialists, who should consider two important drawbacks of the proposed method. First, the amount of primer-dimers is variable and depends on numerous factors such as time and temperature before starting PCR cycling, Mg2+ concentration, the presence of reverse transcriptase, which make sample-to-sample comparison misleading. Second, primer-dimers greatly reduce the yield of PCR. In order to limit the formation of primer-dimers, numerous protocols and enzyme formulations (hot starts) are available.

Contrary to Halford's statement on the limit of template detection (300–1,000), our personal experience and that of many others is that the theoretical limit (one molecule) is easily achieved when reliable hot-start protocols are used. Avoiding the formation of primer-dimer artifacts and the use of internal standards are still the prerequisites for PCR when a sensitive and accurate quantification is required.