Evolution of pathogens reconstructed from ancient human samples

Summary Infectious diseases caused by viral and bacterial pathogens have affected humans throughout history as exemplified by well-known epidemics, such as the Black Death and the Spanish Flu. Still today, many viruses and multi-drug resistant bacterial strains pose a serious threat to public health and the global economy. To overcome this threat, it is not only necessary to analyze the genomes of recent pathogenic strains but to look back in time. Present-day strains can yield only snapshots of their current genomic diversity and allow only indirect inferences about their long-term evolution. However, with ancient DNA research, it is possible to directly detect and trace changes in the genomic make-up of pathogen progenitors through time, thus providing robust historical and evolutionary evidence e.g. for changes in virulence and accurate estimation of long-term mutation rate. In this thesis, I conducted a large-scale bioinformatic screening of shotgun sequencing libraries for the presence of pathogens. The libraries were constructed from DNA extracts of ancient human remains almost entirely from Europe (5000 BCE - 20th century CE). Because infection-induced lesions on human remains are rare and often unspecific I performed this screening without any underlying particular hypothesis and did not limit it to a single pathogen. In total, I analyzed 2255 datasets from 1881 individuals with several pathogen-specific pipelines that were individually adapted from publicly available software. The datasets were investigated for their metagenomic content. Any hits indicating the presence of specific pathogens led to the analysis of the respective genomes. Using this approach, I detected five different pathogens. This thesis particularly looks at three of them, Yersinia pestis (Y. pestis), Hepatitis B virus (HBV) and parvovirus B19 (B19). I identified and reconstructed a Neolithic Y. pestis genome, which helped to reliably estimate the split between the Y. pestis clade and its progenitor Y. pseudotuberculosis. Most of the virulence factors in this genome carried the ancestral states, suggesting a milder disease course in prehistory than in the Middle Ages when infections were caused by more virulent Y. pestis strains. In a Y. pestis genome from the 17th century CE, I detected a depletion of the pla virulence factor compared to previous Black Death strains. This finding prompted me to investigate the pla copy numbers in published Y. pestis genomes. Overall, 15 genomes representing post-Black Death strains (after the 14th century CE) showed this depletion. The decrease of pla might, together with other factors, have contributed to the disappearance of the plague from Europe. Furthermore, I assembled 12 HBV genomes ranging from the 5th millennium BCE to the Middle Ages. Of these, 2 presented as hybrids between present-day human and non-human primate HBV genomes. However, it is unclear if these strains were the progenitors of nowadays HBV diversity or a result of recombination from an unknown ancient HBV diversity. These hybrids were only found in remains older than 3800 years and no longer exist today. For decades, it has been hypothesized that HBV transmissions between humans and non-human primates took place. The hybrids were thus the first authentic evidence for such transmissions in vivo. The other 13 analyzed genomes dated to the Middle Ages reflect the current European HBV diversity. The research conducted helps to better understand the history, distribution, long-term evolution and genetic properties of these pathogens. It expanded the temporal boundaries of ancient DNA pathogen research by reconstructing viral genomes up to 7000 years old and shows that even minute traces can be worthwhile chasing