Introduction
Rapid and cost-effective biodiversity surveillance is a critical, burgeoning tool in ecology and conservation. DNA metabarcoding, or high-throughput DNA sequencing of taxonomically informative barcode genes, has become a well-established tool to efficiently detect many species from bulk DNA samples for biodiversity surveillance. Such sequencing is now frequently used to catalog species from the DNA that organisms shed into the environment (eDNA) (Bohmann et al., 2014). For example, aquatic vertebrate or invertebrate communities can be detected with eDNA metabarcoding of water samples (Shaw et al., 2016; Stat et al., 2017; Valentini et al., 2016), animal diets can be detected from eDNA metabarcoding of feces (De Barba et al., 2014; Massey et al., in press; Roffler et al., 2021), and plants and fungi can be detected by eDNA metabarcoding of pollen or ice core samples (Bell et al., 2019; Kraaijeveld et al., 2015; Varotto et al., 2021). Similarly, an emerging technique to monitor the diversity of terrestrial vertebrates uses animal-feeding invertebrates (whether feeding upon blood, flesh, or scat) as direct sources of species’ DNA and is called invertebrate-derived DNA (iDNA) (Batovska et al., 2018; Calvignac-Spencer et al., 2013; Drinkwater et al., 2019, 2020; Gogarten et al., 2020; Kocher et al., 2017; P. S. Lee et al., 2016; Schnell et al., 2012; van der Heyde et al., 2020; D. W. Yu et al., 2012). Similar to camera traps, a conventional method for terrestrial vertebrate surveillance, metabarcoding blood, carrion, or fecal meals from iDNA samplers can be utilized as a non-invasive and efficient method at profiling a diversity of species in often remote or dense habitats. However, these methods differ in that camera trapping is strongly biased for the detection of medium to large terrestrial mammals, with lower detection probabilities of arboreal species, smaller bodied species, and other non-mammal taxa (Burton et al., 2015; O’Brien & Kinnaird, 2011). These drawbacks could make iDNA metabarcoding a more attractive method for biodiversity monitoring, particularly if iDNA is better able to detect the birds, amphibians, reptiles, small mammals, and arboreal mammals that camera trapping often fails to detect.
Leeches (Drinkwater et al., 2019; Schnell et al., 2012, 2018; Weiskopf et al., 2018) and carrion flies (Calvignac-Spencer et al., 2013; P. S. Lee et al., 2016; Lynggaard et al., 2019; Rodgers et al., 2017) have been the primary sources of validating iDNA metabarcoding for monitoring vertebrate biodiversity. The limited number of studies utilizing metabarcoding with leeches or carrion flies as iDNA samplers have shown that both taxa effectively profile biodiversity in remote or difficult to access environments such as tropical forests (Drinkwater et al., 2020; Gogarten et al., 2020; P. S. Lee et al., 2016; Weiskopf et al., 2018). There is a growing literature of comparing carrion flies as iDNA sources for metabarcoding to camera trapping (Gogarten et al., 2020; P. S. Lee et al., 2016; Rodgers et al., 2017) but even with their utility for profiling animal diversity, neither leeches nor carrions flies are widely collected for other research purposes.
In contrast, some hematophagous insects such as mosquitos and sandflies are important pathogen vectors worldwide and are routinely sampled for disease surveillance. Should these pathogen vectors be effective samplers of vertebrate biodiversity, then using iDNA to profile biodiversity would be an attractive supplemental and/or complementary pursuit to other study objectives using vector species (Batovska et al., 2018; Kocher et al., 2017). We are aware of only one study that has compared mosquitos and sandflies as iDNA samplers and found that both were seemingly effective a describing vertebrate diversity (Kocher et al., 2017). However, these results were not compared to data from a more established iDNA sampler or to more conventional methods of diversity surveillance.
A direct comparison of the emerging iDNA samplers (vectors such as mosquitos and sandflies) to the commonly utilized iDNA taxa groups (such as carrion flies or leeches) and finally to traditional methods of biomonitoring (such as camera traps) is needed to evaluate the diversity revealed by each iDNA source. Such a comparison will set the groundwork for assessing common biases, benefits, and shortfalls associated with each method given the current lack of comparison of established methods to the emerging iDNA samplers of mosquitos and sandflies. Invertebrate taxa differ in their feeding ecologies and other life history traits, which can introduce taxonomic biases into biodiversity assessments when these taxa are used as iDNA sources particularly when utilizing these methods with small sample sizes or spatial scales. Thus, this comparison is particularly needed at a spatial scale where these methods can be evaluated for both alpha and gamma diversity given the same suite of available species.
Here, we aim to assess the effectiveness of iDNA metabarcoding for biodiversity surveillance using landscape scale, bulk sampling of insect groups including carrion flies, sandflies, and mosquitos. We first compare the three iDNA samplers to camera trapping (the established method of vertebrate diversity surveillance) to determine if the two methods reveal similar species richness, composition, and relative abundance across the sampled landscape. Second, we compare the biodiversity found with each of the iDNA samplers to assess the efficiencies and potential biases with using different sources of iDNA as biomonitoring tools. To achieve these comparisons, we examine both gamma and alpha diversity with each iDNA sampler and tease apart the likely causes of dissimilarity.