Single base errors in PCR products from avian museum specimens and their effect on estimates of historical genetic diversity

Conservation genetic studies often employ DNA extracts from museum specimens for comparisons with extant populations to monitor temporal changes in genetic diversity. Here, we report on artifact base changes in mitochondrial DNA sequences amplified from relatively recent (<= 35 years) museum specimens of indigobirds (Vidua spp.). Single base errors were confirmed by replicate sequencing and included both double peaks and artifact substitutions at rates of similar to 3 x 10(-4) and similar to 1 x 10(-4) per base- pair, respectively, resulting in one or more errors or ambiguities in an 1100 base pair sequence in 21% of 219 samples. Most errors involved C -> T changes on the L-strand, presumably due to deamination of cytosine in the template. The error rates encountered here bias comparisons of haplotype number between historical and extant populations, such that the `loss' of artifact haplotypes present in a historical sample could be incorrectly attributed to a population decline or bottleneck. Sequencing errors due to miscoding lesions in template DNA have so far been reported only from ancient and formalin-fixed tissue, but they may also affect relatively recent museum samples, as shown here, and perhaps also non-invasive samples that typically yield low-quality DNA.