Understanding the diet of an endangered species illuminates the animal’s ecology, habitat requirements, and conservation needs. However, direct observation of diet can be difficult, particularly for small, nocturnal animals such as the Pacific pocket mouse (Heteromyidae: Perognathus longimembris pacificus). Very little is known of the dietary habits of this federally endangered rodent, hindering management and restoration efforts. We used a metabarcoding approach to identify source plants in fecal samples (N = 52) from the three remaining populations known. The internal transcribed spacers (ITS) of the nuclear ribosomal loci were sequenced following the Illumina MiSeq amplicon strategy and processed reads were mapped to reference databases. We evaluated a range of threshold mapping criteria and found the best-performing setting generally recovered two distinct mock communities in proportions similar to expectation. We tested our method on captive animals fed a known diet and recovered almost all plant sources, but found substantial heterogeneity among fecal pellets collected from the same individual at the same time. Observed richness did not increase with pooling of pellets from the same individual. In field-collected samples, we identified 4–14 plant genera in individual samples and 74 genera overall, but over 50 percent of reads mapped to just six species in five genera. We simulated the effects of sequencing error, variable read length, and chimera formation to infer taxon-specific rates of misassignment for the local flora, which were generally low with some exceptions. Richness at the species and genus levels did not reach a clear asymptote, suggesting that diet breadth remained underestimated in the current pool of samples. Large numbers of scat samples are therefore needed to make inferences about diet and resource selection in future studies of the Pacific pocket mouse. We conclude that our minimally invasive method is promising for determining herbivore diets given a library of sequences from local plants.
- Learn More: https://doi.org/10.1371/journal.pone.0165366
- USGS Source: Publications Warehouse (indexId: 70191088)