Disease and pathogens have affected human populations throughout history, something that the global pandemics of the 21^st century can attest for. With the development of methods for DNA extraction and sequencing during the last decade, it is now possible to study ancient pathogen evolution and transmission more in depth, within the field of ancient metagenomics. However, a long-standing challenge in ancient metagenomics has been high error rates and false positive identifications. In this thesis, I have aimed to initially improve the methods for analysing ancient DNA data, and further to study the presence and evolution of pathogens in populations across human prehistory. In chapter I, I present aMeta, an accurate ancient metagenomics profiling workflow that has been designed to minimize the number of false positive identifications, as well as to streamline computer memory usage. Using simulated as well as authentic ancient DNA data, aMeta was benchmarked against an existing workflow, and its superior sensitivity and specificity in both microbial detection and authentication was demonstrated. Further, we could show its substantially lower usage of computer memory. In chapter II, the aMeta workflow was applied on a dataset consisting of 38 individuals from four Mesolithic and Neolithic Scandinavian human cultural complexes. Several species of bacteria were identified in the dataset, for example the bacterium Salmonella enterica in two individuals from the Battle Axe cultural complex. Since osteological examination did not present any physical damage to the bones, this disease may have been the cause of death for the infected individuals. Several species of the bacterial genus Yersinia were identified in individuals from the Funnel Beaker culture context, and denser populations in an agricultural context may have facilitated the transmission of these pathogens. Further, in Mesolithic and Neolithic hunter-gatherers, two pathogenic species of the genus Neisseria were identified, representing the, to our knowledge, earliest findings of the species to date. In chapter III, aMeta was applied to a dataset from Mexico, consisting of 41 individuals dated between 900 – 1800 CE. In one individual, we identified DNA from the bacterium Vibrio cholerae, the causing agent of cholera. We created a phylogeny consisting of available, globally collected Vibrio genomes and concluded that our finding, the earliest of V. cholerae to date, likely belongs to a non-choleric strain and thus may not have been the cause of an epidemic. Further, the finding indicates that cholera may have arrived in the Americas decades earlier than previous research has shown. In chapter IV, we presented genomic data from 40 individuals in northeast Asia, dated between circa 16,900 and 550 years ago. Population demographics showed genetic affinity between the analysed individuals and present-day human populations in Asia and Native America. We further used the metagenomics tool Malt to identify Yersinia pestis reads in two individuals from 4,400 and 3,800 years ago respectively, representing the most northeastern ancient finding of the bacterium.