Improved assemblies and comparison of two ancientYersinia pestisgenomes

ABSTRACT Yersinia pestis is the causative agent of the bubonic plague, a disease responsible for several dramatic historical pandemics. Progress in ancient DNA (aDNA) sequencing rendered possible the sequencing of whole genomes of important human pathogens, including the ancient Yersinia pestis strains responsible for outbreaks of the bubonic plague in London in the 14th century and in Marseille in the 18th century among others. However, aDNA sequencing data are still characterized by short reads and non-uniform coverage, so assembling ancient pathogen genomes remains challenging and prevents in many cases a detailed study of genome rearrangements. It has recently been shown that comparative scaffolding approaches can improve the assembly of ancient Yersinia pestis genomes at a chromosome level. In the present work, we address the last step of genome assembly, the gap-filling stage. We describe an optimization-based method AGapEs (Ancestral Gap Estimation) to fill in inter-contig gaps using a combination of a template obtained from related extant genomes and aDNA reads. We show how this approach can be used to refine comparative scaffolding by selecting contig adjacencies supported by a mix of unassembled aDNA reads and comparative signal. We apply our method to two data sets from the London and Marseilles outbreaks of the bubonic plague. We obtain highly improved genome assemblies for both the London strain and Marseille strain genomes, comprised of respectively five and six scaffolds, with 95% of the assemblies supported by ancient reads. We analyze the genome evolution between both ancient genomes in terms of genome rearrangements, and observe a high level of synteny conservation between these two strains.