Advancement to hair-sampling surveys of a medium-sized mammal: DNA-based individual identification and population estimation of a rare Australian marsupial, the spotted-tailed quoll (Dasyurus maculatus)

Context. Enumeration of cryptic/rare or widely distributed mammal species is exceedingly difficult for wildlife managers using standard survey methods. Individual identification via non-invasive hair-DNA methods offers great promise in extending the information available from hairs collected to survey for presence/absence of a species. However, surprisingly few wildlife studies have attempted this because of potential limitations with the field method and genetic samples. Aim. The applicability of hair DNA to identify individuals and estimate numbers was assessed for a rare, medium-sized Australian marsupial carnivore, the spotted-tailed quoll (Dasyurus maculatus). Methods. Hair samples were obtained remotely in the field with baited hair-sampling devices (known as handi-glaze hair tubes) that permit multiple visitations by individuals and species. A hierarchical approach developed and applied to the DNA extraction and PCR protocol, based on single and four pooled hairs of each collected sample, was used to assess genotype reliability (cross-species DNA mixing, allelic dropout and false allele errors) and enumerate the local study population. These results were compared against a concurrent live-cage trapping survey that was equivalent in scale and trap density to enable a rigorous evaluation of the efficiency and reliability of the DNA-based hair-sampling technique. Key results. Of the 288 hair devices deployed, 52 (18%) captured spotted-tailed quoll hair and the majority (90%) of these samples provided adequate DNA to genetically profile individuals at 10 microsatellite loci and a sexing marker. The hierarchical approach provided a feasible way to verify whether cross-species DNA mixing had occurred in the pooled-hair DNA extracts by comparing the results against the independent single-hair DNA extract, and assess genotyping reliability of both DNA concentrations. Fewer individuals were detected using hair-sampling (n = 16) than live-trapping (n = 21), despite hair-sampling occurring over a longer period (40 cf. 26 nights). Conclusions. The population-level information gained by the DNA-based technologies adds considerable value to the remote hair-sampling method which up until the present study had been used to detect the presence of medium-sized mammals. Our study demonstrated the utility of the DNA-based hair-sampling method to identify spotted-tailed quoll individuals and for surveying local populations. However, improvements to the hair-sampling method, such as increasing the density of stations or the provision of a food reward, should be considered to enhance sampling efficiency to allow the enumeration of local populations. Implications. The use of remote hair-sampling devices that permit multiple visitations and do not require daily collection can be feasible and reliable to genetically identify individuals when coupled with appropriate strategies. By combining single- and pooled-hair DNA extracts, a good compromise between laboratory efficiency and data integrity is afforded.