Host whole genome sequence data represent an untapped resource for characterising affiliated parasite diversity

Parasites are ubiquitous and exert varied ecological and evolutionary pressures on their hosts. Yet, characterising parasite diversity and distributions can be challenging and costly. Leveraging existing data to identify parasites is thus an attractive alternative. High-throughput sequencing (HTS) can generate whole genome sequence (WGS) data which are increasingly freely available in public repositories and represent an untapped resource for characterising parasites affiliated with hosts. In this study, we examine WGS data generated for the silvereye (Zosterops lateralis), to identify endogenous eukaryotic parasites that were inadvertently captured during host sequencing. We compared detection of parasite genera by this approach with detection via 18S metabarcoding. Mining WGS data for parasite DNA revealed the broadest range of genera. Results were verified by traditional microscopy of blood slides and conducting a targeted multiplex Polymerase Chain Reaction (PCR) for haemosporidian parasites. Detection of haemosporidians was largely consistent across microscopy, multiplex PCR and WGS data while 18S metabarcoding entirely failed to detect this group of parasites. Our results demonstrate that existing WGS datasets can be used to estimate endoparasite diversity and provide greater insights on diversity than metabarcoding whilst also avoiding the costs and challenges of direct sampling. We provide a framework outlining opportunities and constraints to consider when mining WGS data to identify parasite sequences. The framework particularly stresses the influences of sequencing depth, database completeness, and methodological biases. Our findings demonstrate how repurposing existing WGS data can provide a cost-effective and informative means of unravelling complex host-parasite interactions in future disease ecology studies.