The southern meadow jumping mouse (Zapus luteus luteus) is Endangered in the US due to loss of riparian habitat and knowledge of diet can focus conservation efforts. Context and life stages of diet have been documented for other species of Zapus (seeds, insect larvae, and hypogeous fungi), possibly shifting toward greater diversity in seed predations prior to hibernation. Yet, challenges in field observation, thorough digestion, and lack of scalability have hindered a thorough survey of dietary taxonomies. We collected the feces of 165 Z. l. luteus from across the geographic range. Emphasizing granivorous diet, we used DNA metabarcoding to resolve its dietary taxonomies. With negligible impact of field contamination, we found a varied diet most frequently of graminoids, forbs, and lepidoptera. Detecting up to nine dietary taxa in individuals, Z. l. luteus consumed more and different combinations of plant taxa as they approached the known hibernation window. Our work emphasizes the importance of forbs and graminoids as the foundation of the diet, providing primary sustenance and habitat for insect prey. The breadth of different diet items they consume suggests plasticity in resource use, potentially accommodating diverse patterns of seed production throughout their active period. We also hypothesize that a peak in seed availability in the late season could play a role in their accumulation of pre-hibernation fat stores.

We PCR amplified five DNA metabarcode markers (plant ITS2, plant rbcL, arthropod COI, fungal ITS2, and rodent cytochrome B) among two PCR multiplex reactions from the fecal DNA of Zapus luteus luteus. Libraries were prepared with strict non-combinatorial indexing and sequenced on an Illumina MiSeq V3 600 cycle kit (SRA BioProject accession: PRJNA833142). We demultiplexed samples and markers and processed data for marker datasets separately. Briefly we denoised and merged paired-end reads, post-clustered plant markers for diversity analysis, used a 1.1% relative read abundance filter to avoid exogneous contaminants and spurious features, filtered to exclude off-target amplifications within each marker, and classified using Naive Bayes classifiers. We used the cytochrome B marker to verify that the fecal sample belonged to Zapus luteus luteus (classified using BLAST alignment to NCBI and least common ancestor classification). Any fecal sample with the identifying DNA of other sympatric rodents was excluded from the dataset. We also sequenced bait used for live capture and excluded any features found in the bait from the entire dataset. Taxonomies were cross-referenced to a local database of plants available at capture localities or state-wide and national databases, and then collapsed a classification to a higher taxonomic level if the taxa were not known to occur there. Of the fungal taxonomies, we only retained hypogeous fungal taxa based on our study objectives. Both collapsed-filtered and annotated-unfiltered versions of the dataset are available. Diversity analyses (alpha and beta diversity) are limited to two study areas based on consistency in trapping and fecal sampling. Both diversity datasets have been rarified to 2500 reads following rarefaction analysis.

R scripts are available for statistical analyses of alpha and beta diversity. Datasets used as input and analysis procedures are noted at the beginning of and throughout the scripts. Please see README_file.txt for more information regarding the R scripts or contents of the datasets.