Non-random DNA damage resulting from heat treatment: implications for sequence analysis of ancient DNA

Abstract

To mimic damage occurring during preservation of ancient DNA, we heated wheat DNA at 95 °C for 2–21 days. Samples were amplified by PCRs directed at a 181-bp segment of the mitochondrial atpA gene. PCR products were cloned and multiple sequences obtained. We anticipated that nucleotide changes indicating damage would be randomly distributed, but this was not the case. Of the 124 cloned sequences, 107 displayed the same adenine to guanine change, this change being seen in clones from all of the eight heated samples. Fifteen of the cloned sequences, from six samples, showed the same apparent deletion at position 92. If non-random sequence changes are a widespread feature of DNA degradation then current strategies for analysis of ancient DNA sequences must be revised, as these could lead to nucleotide variations that are artefacts resulting from DNA diagenesis being mistaken for genuine sequence features of the ancient DNA.