Invertebrates for vertebrate biodiversity monitoring: comparisons using three insect taxa as iDNA samplers

Abstract

Metabarcoding of environmental DNA (eDNA) is now widely used to build diversity profiles from DNA that has been shed by species into the environment. There is substantial interest in the expansion of eDNA approaches for improved detection of terrestrial vertebrates using invertebrate-derived DNA (iDNA) in which hematophagous, sarcophagous, and coprophagous invertebrates sample vertebrate blood, carrion, or feces. Here, we use metabarcoding and multiple iDNA samplers (carrion flies, sandflies, and mosquitos) to profile gamma and alpha diversity in a dry, tropical forest in the southern Amazon. Our main objectives were to (1) compare diversity found with iDNA to camera trapping, which is the conventional method of vertebrate diversity surveillance and (2) compare each of the iDNA samplers to assess the effectiveness, efficiency, and potential biases associated with each sampler. Carrion flies were the most effective sampler, despite the least amount of sampling effort and the fewest number of individuals captured for metabarcoding, in describing vertebrate biodiversity followed by sandflies. Camera traps had the highest median species richness at the site-level but showed strong bias towards carnivore and ungulate species and missed much of the diversity described by iDNA methods. Mosquitos showed a strong feeding preference for humans as did sandflies for armadillos, thus presenting potential utility to further study related to host-vector interactions.

Phylogenetic and biogeographical traits predict unrecognized hosts of zoonotic leishmaniasis

“…While our follow up analyses do not indicate that model results are biased by uneven study effort, this finding may suggest that zoonotic Leishmania hosts are generally under sampled. Though leishmaniasis is a neglected tropical disease primarily impacting resource-poor communities, state-of-the-art zoonotic Leishmania diagnostic and surveillance techniques are available [16,57,[65][66][67][68][69]. As such, under sampling Leishmania hosts primarily stems from both the lack of funding to employ these tools during large scale surveillance efforts and the large number of animal species that are potentially hosts.…”

“…Notably, [56] used machine learning to predict unrecognized vectors of leishmaniasis. Metabarcoding of sandflies and their blood meals [66,81] may be useful in linking model predictions of vectors and hosts to local transmission cycles.…”

Medium‐ to large‐bodied mammal surveys across the Neotropics are heavily biased against the most faunally intact assemblages

“…This is a persistent sampling bias in the Neotropics (e.g. Oliveira et al 2017), despite the rapid emergence of modern non-invasive sampling methods, such as eDNA, metagenomics, camera trapping, and automated soundscapes that do not require extended periods of fieldwork (Massey et al 2021).…”