The Impact of Reference Genome Divergence on Ancient DNA Damage Detection in Metagenomic Contexts

Abstract

The reliability of ancient DNA (aDNA) authentication depends on detecting characteristic damage patterns, particularly cytosine deamination at fragment ends. However, in ancient metagenomic studies, sequence divergence between aDNA reads and available reference genomes may obscure such damage signals. We systematically evaluated how reference genome divergence, read count, read length, and damage levels affect aDNA damage profiles using both empirical datasets and controlled simulations. Using ancient Yersinia pestis and Hepatitis B virus data, we show that mapping to divergent reference genomes significantly reduces the detectability and intensity of characteristic damage patterns, particularly at low read counts. Simulations further revealed that reference genome identity is the strongest predictor of damage intensity, while read count primarily influences damage stochasticity. We introduce a correction matrix that adjusts C-to-T damage profiles for reference divergence, improving damage signal recovery. Our findings highlight methodological considerations for authenticating aDNA in metagenomic contexts, particularly when closely related reference genomes are unavailable.