At best, most ancient samples contain no or only small amounts of amplifiable endogenous DNA. This, combined with a complex and poorly understood contamination risk in ancient DNA studies, involves a high risk of false-positive results (Cooper and Poinar 2001; Hofreiter et al. 2001b; Marota and Rollo 2002; Willerslev et al. 2004b; Pâabo et al. 2004; Willerslev and Cooper 2005). Traditional contamination is separated into laboratory and sample contamination.
To avoid laboratory contamination, all pre-PCR work should be carried out in dedicated isolated ancient DNA facilities with separate ventilation systems, nightly UV irradiation, and positive air pressure. The work should be carried out following strict protocols with bodysuits, facemasks, and gamma-sterilized gloves (Hebsgaard et al. 2005; Willerslev and Cooper 2005). Blank-extraction and PCR-amplification controls should be incorporated. Blank controls cannot by themselves guarantee detection of laboratory contamination, due to the sporadic nature of contamination and carrier effects (Cooper and Poinar 2001; Marota and Rollo 2002; Cooper 1993; Handt et al. 1994; Pââbo et al. 2004; Willerslev and Cooper 2005).
Another risk of contamination is carryover of PCR products, which can lead to high levels of amplicons rapidly spreading through laboratories, making it easy to obtain false-positive amplification products (Willerslev and Cooper 2005).
It is impossible to discount minor amounts of laboratory-based contamination, even for the most comprehensive laboratory setup. This holds especially true in human and microbial studies due to the universal distrinution of these organisms in laboratory settings (Rollo and Marota 1999; Willerslev et al. 2004b; Pââbo et al. 2004).
However, high contamination risk can also be applied to studies of rare organisms (even extinct species) if close modern relatives are processed in the same laboratory or large amounts of amplicons are produced, such as in large-scale genetic population studies (Shapiro et al. 2004). Fortunately, laboratory contamination, although a serious concern, can be detected by the following simple authentication criteria (Pââbo 1989; Cooper and Poinar 2001; Hebsgaard et al. 2005). The independent replication of results by another laboratory is the strongest argument against laboratory contamination, because it is unlikely that the same contaminant sequence will be independently sequenced in another laboratory.
Much more challenging is sample contamination, because it is much more difficult to exclude. In most human and microbial studies there is currently no way to clearly distinguish an endogenous DNA sequence or culture from that of a contaminant (Rollo and Marota 1999). The problem is especially pronounced where the samples have been handled by several individuals during excavation. In the same way, microbes can easily contaminate samples just by passive or active movement. Even microbes known to be associated with a particular specimen might have unknown relatives or even identical ecotypes in the surrounding environment (Gilbert et al. 2005a). Sample contamination can only be excluded for sequences obtained from morphologically identifiable specimens, with restricted extant distributions and well-known diversity (e.g., many vertebrates and some higher plants), though recent sequencing of DNA directly from sediments or ice (Willerslev et al. 2003, 2007; Johnson et al. 2007) complicates authentication for these groups.
Was this article helpful?