Appendix S3: Supplemental methods and discussion
Methods
We assigned each OTU to one of 4 categories based on similarity scores of the top 100 matches as follows:
1a: true species match, >99% similarity to a single species in local assemblage
1b: likely species match, >98% similarity to single species in local assemblage
2: >98%, matches multiple species but only one in local assemblage
3: >98%, matches multiple species in the same genus in local assemblage
4: >98%, matches multiple genera in the same family in local assemblage
For the purposes of this study, the local assemblage was considered to include any arthropod with records in North America. Distribution records were considered from BOLD, BugGuide (BugGuide 2016), or Moth Photographers Group (MPG 2016). For each OTU matching a distinct species (Level 1) we verified migrant, pest, or beneficial status using relevant sources. Matches other than level 1 (species) can be inaccurate unless there is a strong phylogenetic match among database samples (Wilson et al. 2011). Therefore we only accepted matches at genus or family level where the clade comprised members of a single taxon (Burgar et al. 2013).
Records in BOLD are not equally distributed among insect orders (Jinbo et al. 2011). For a less biased estimation of OTU assignment across orders, we used BLASTn to query each OTU against a subset of the NCBI GenBank nucleotide database containing all CO1 sequences, using expected alignment value = 0.00001. We imported the BLASTn results into MEGAN (MEtaGenome Analyzer v5, (Huson et al. 2007)), where they were mapped and visualized against the NCBI taxonomic framework (min. bit score = 1.0, top percentage = 10%, min. support = 1). We mapped OTUs to order based on the following criteria. If MEGAN provided a terminal order that did not conflict with a match to a single species in BOLD, we used the terminal order from MEGAN. If MEGAN did not provide a terminal order, but all matches > 98% in BOLD were to a single order, we assigned the OTU to that order.
Discussion
Earlier studies also document a wide breadth of diet in T. brasiliensis bats, including at least 12 arthropod orders and 38 identified families and seasonal changes in bat diets (Lee and McCracken 2005; McWilliams 2005). Working at Carlsbad Caverns in New Mexico, McWilliams (2005) documented dramatic seasonal shifts that agree with the earlier studies in Texas, with moths representing only 6.1% of species in bat diet in late July, down from a peak of 79.1% in late April and before a return to 60.1% in early September.
Our results also support diet studies on other bats. A recent study of the European free-tailed bat T. teniotis, showed that, on average, 56.9% of prey species in each fecal pellet were migratory moths (Mata et al. 2016). While our study and that of Mata et al. (2016) were on different continents, three of the consumed migratory moth species were the same in both studies (Agrotis ipsilon, Spodoptera exigua, and Peridroma saucia). In North America, migratory hoary bats (Lasiurus cinereus) also were found to consume 99.2% moth species, of which 51.1% were noctuids (Valdez and Cryan 2009).
The insect species consumed by the bats in this study (Table S3) included many of the prey taxa eaten by T. brasiliensis in an earlier study at the same site (Lee 1999). These include crickets (found in all samples in this and the previous study), lygaeid, pentatomid, and cicadellid bugs, scarab beetles, neuropterans, and tephritid flies. There were notable differences between this study and earlier descriptions of T. brasiliensis diet. We found no damselflies (Whitaker et al. 1996; Lee 1999), termites (McWilliams 2005), chrysomelid beetles (Whitaker et al.1996; Lee 1999; McWilliams 2005), or spiders (Whitaker et al. 1996; McWilliams 2005). However, because representation in the BOLD database is not equal across orders (Jinbo et al. 2011), it is possible that some of the taxa missing in this study were due to inadequate DNA database coverage rather than actual absence in the diet.
We found two types of problems while matching OTUs to species. In the first, some OTUs matched to multiple species, probably due to short sequence length. Prey DNA extracted from predator fecal samples is highly degraded and the short sequence lengths represented only a portion of the target CO1 gene used for species identification, resulting in inconclusive matches with reference databases. The relatively low number of OTUs assigned to species reported in this study as compared to that of other studies (e.g. Clare et al. 2011; Mata et al. 2016; Van Den Bussche et al. 2016; Vesterinen et al. 2016) was likely a result of our taking a more conservative approach. While most OTUs matched species at greater than 98%, most also matched to multiple species in North America at that level, and thus we could not assign them to a single species and instead assigned them to a higher category than Level 1.
In the second problem, some species were represented by multiple OTUs, possibly due to high genetic diversity in the reference databases for common or pest species. The 129 species identified represented 25% of all 656 OTUs, and 62 different OTUs mapped to 19 species. Other studies similarly found duplicate OTUs in different species (Mata et al. 2016; Van Den Bussche et al. 2016). Nine of the 19 species accounting for multiple OTUs were known pests. In some instances OTUs appearing rarely in the dataset (10 sampling days or fewer) matched to the same species as very common OTUs. Thus until it is possible to match most OTUs to species, patterns in prey communities should be interpreted with caution.
Two additional issues were identified in this analysis that should be considered while interpreting results. First, the most common OTUs were associated with two species of crickets (Gryllus rubens and G. veletis) that are known to be pests, and G. rubens is also considered to be migratory. OTUs appearing in most or all samples should be evaluated as possible contaminants, and those OTUs also appeared in some extraction blanks. However, crickets were very common in morphological diet studies in the area (Lee 1999) and have been caught in aerial nets in southern Texas during autumn (Krauel, unpublished data). To be conservative, crickets were not included in further diet analysis in this study. Finally, we interpreted the number of OTUs found on each sampling day as a measure of species richness. Three sampling days had fewer than 50 OTUs found and it is possible that these values represent a sequencing artifact rather than actual low richness on those days, although we found no evidence of low yield during extraction or sequencing of those samples.
References
BugGuide (2016). Iowa State University Dept of Entomology. http://bugguide.net.
Burgar, J.M., Murray, D.C., Craig, M.D., Haile, J., Houston, J., Stokes, V. & Bunce, M. (2013) Who's for dinner? High-throughput sequencing reveals bat dietary differentiation in a biodiversity hotspot where prey taxonomy is largely undescribed. Molecular Ecology, doi: 10.1111/mec.12531.
Clare EL, Barber BR, Sweeney BW, Hebert PDN, Fenton MB (2011) Eating local: influences of habitat on the diet of little brown bats (Myotis lucifugus). Mol Ecol 20:1772-1780. doi:10.1111/j.1365-294X.2011.05040.x
Huson DH, Auch AF, Qi J, Schuster SC (2007) MEGAN analysis of metagenomic data. Genome Research 17 (3):377-386. doi:10.1101/gr.5969107
Jinbo, U., Kato, T. & Ito, M. (2011) Current progress in DNA barcoding and future implications for entomology. Entomological Science, 14, 107-124.
Krüger F, Clare EL, Symondson WOC, Keišs O, Pētersons G (2014) Diet of the insectivorous bat Pipistrellus nathusii during autumn migration and summer residence. Mol Ecol 23 (15):3672-3683. doi:10.1111/mec.12547
Lee, Y.F. (1999) Resource use and foraging activity of Mexican free-tailed bats Tadarida brasiliensis mexicana (Molossidae). University of Tennessee.
Lee, Y.F. & McCracken, G.F. (2005) Dietary variation of Brazilian free-tailed bats links to migratory populations of pest insects. Journal of Mammalogy, 86, 67-76.
Mata, V.A., Amorim, F., Corley, M.F.V., McCracken, G.F., Rebelo, H. & Beja, P. (2016) Female dietary bias towards large migratory moths in the European free-tailed bat (Tadarida teniotis). Biology Letters, 12, 5.
McWilliams, L.A. (2005) Variation in diet of the Mexican free-tailed bat (Tadarida brasiliensis mexicana). Journal of Mammalogy, 86, 599-605.
MPG (2016) North American Moth Photographers Group. http://mothphotographersgroup.msstate.edu/. Mississippi Entomological Museum, Mississippi State University.
Ratnasingham S, Hebert PDN (2007) BOLD: The Barcode of Life Data System (www.barcodinglife.org) Mol Ecol Notes 7:355-364 doi:10.1111/j.1471-8286.2006.01678.x
Valdez, E.W. & Cryan, P.M. (2009) Food habits of the hoary bat (Lasiurus cinereus) during spring migration through New Mexico. Southwestern Naturalist, 54, 195-200.
Van Den Bussche, R.A., Lee, D.N., Judkins, M.E., Dyer, J.E., Thompson, D.M., Stark, R.C., Puckette, W.L. & Fuller, B. (2016) Molecular dietary analysis of the endangered Ozark big-eared bat (Corynorhinus townsendii ingens). Acta Chiropterologica, 18, 181-191.
Vesterinen, E.J., Ruokolainen, L., Wahlberg, N., Peña, C., Roslin, T., Laine, V.N., Vasko, V., Sääksjärvi, I.E., Norrdahl, K. & Lilley, T.M. (2016) What you need is what you eat? Prey selection by the bat Myotis daubentonii. Molecular Ecology, 25, 1581-1594.
Whitaker, J.O., Neefus, C. & Kunz, T.H. (1996) Dietary variation in the Mexican free-tailed bat (Tadarida brasiliensis mexicana). Journal of Mammalogy, 77, 716-724.
Wilson, J.J., Rougerie, R., Schonfeld, J., Janzen, D.H., Hallwachs, W., Hajibabaei, M., Kitching, I.J., Haxaire, J. & Hebert, P.D.N. (2011) When species matches are unavailable are DNA barcodes correctly assigned to higher taxa? An assessment using sphingid moths. BMC Ecology, 11, 18.