Fecal samples of seed-dispersing Neotropical birds for plant metabarcoding

Fruit-eating birds drive seed dispersal in naturally recovering tropical ecosystems, ultimately shaping the trajectory of forest regeneration. Novel molecular techniques can offer a powerful pathway to analyze bird diet using feces samples and inform seed dispersal dynamics. The detection of plant DNA in bird feces using a metabarcoding approach can complement the information about dispersal in areas where seed dispersal events are infrequent, as well as identify ingested seeds collected in hyper-diverse tropical regions, where seed reference collections for morphological identification are challenging to curate. In this study, we applied a novel protocol to detect plant DNA in bird feces and compared the plant genera identified using metabarcoding with the seeds morphologically identified in the same bird droppings. We extracted plant DNA from bird droppings and amplified a universal mini-barcode from the rbcL gene region. To reduce the costs, samples were molecularly tagged and pooled in-house prior to metabarcoding sequencing. Plant genera were identified by comparing the obtained sequences with online reference databases and cross-referencing local plant species lists. Also, seeds found in droppings were morphologically identified using seed reference collections. We report a high success rate of amplification and sequencing of plant DNA obtained from bird droppings. However, failure to detect DNA of some seeds present in the dropping samples, suggesting that the metabarcoding approach is complementary to established seed identification methods and complement the diet information. Molecular analyses of animal feces provide robust diet information, but may be difficult to apply to seed dispersal studies due to the uncertainty of whether dispersal occurred. Also, metabolites produced by plants may lead to taxon-specific biases when conducting community level analyses. While a metabarcoding approach is a promising tool, the evaluation of incongruencies between molecular analyses and traditional methods is crucial for further application in seed dispersal studies.