Abstract: Next-generation sequencing (NGS) benefits to entomological and disease surveillance have been well-proven, yet the two approaches are rarely combined. Here, we describe the development and application of a metabarcoding approach to vector and arbovirus surveillance, including the identification of vertebrate hosts, to a large sample of light trap collections from Southeast Queensland, Australia. Virus surveillance was performed using a combination of pan-Alphavirus, pan-Flavivirus, and Ross River virus (RRV) specific primer sets. The resulting amplicons were combined and indexed with those targeting invertebrates (COI and ITS2) and vertebrate blood-meal hosts (CytB) prior to sequencing to reduce library preparation time and expense. DNA sequencing confirmed the presence of all morphologically identified mosquito species as well as commonly misidentified species and difficult to differentiate biting midge species (Culicoides). The general agreement between the proportion of NGS reads and individual species abundance allows users to reliably infer the overall composition of bulk light trap collections without morphological identification. Importantly, identification of a variety of vertebrate blood-meal hosts was observed, including humans and a variety of known reservoir host species for RRV and other endemic diseases, in the absence of visibly blood-fed females. Lastly, a diversity of insect-specific and medically important viruses, including RRV, was detected. Overall, metabarcoding shows great promise as an efficient vector and arbovirus surveillance tool and means of elucidating spatially and temporally variable vector-host relationships.
