COS 85-5
Molecular identification of the wood thrush diet and validation of a protocol for studies of avian diets
DNA barcoding markers and next-generation sequencing (NGS) show great promise in applications to food web studies. By extracting mixed community DNA from an animal’s scat or stomach contents and then using polymerase chain reaction (PCR) and NGS with public reference databases, this method can identify prey species DNA while overcoming limitations of traditional foraging studies. For example, field observations are both time and labor intensive and biased by the detectability of focal species. Likewise, morphological identification of organisms in scat and stomach contents are labor intensive, require expert knowledge of prey morphology, and can be biased by differences in digestibility of prey. We tested the efficacy of genetic methods to quantify diet diversity and breadth for a declining neotropical migratory bird, the wood thrush (Hylocichla mustelina). We compare species visually detected in samples to those found with an NGS molecular method. We then compare molecular and morphological results from scat samples and stomach contents from the same individuals, to determine how much information is lost during digestion. Three COI barcoding gene primer sets are tested to inform protocols. The success of this approach strongly depends on data generated by universal molecular markers and if information is lost during digestion.
Results/Conclusions
Dissections of partially digested invertebrate prey in scats provided morphological evidence of several orders: Lepidoptera, Coleoptera, Hemiptera, Orthoptera, Diplopoda, and Araneae. Some taxa were identifiable to lower levels. Molecular fingerprints, a community proxy, suggest that different PCR primers give different sample species abundance but mean species amplified is similar between primers (F2,96= 0.69, p=0.51). Pilot sample cloning and Sanger sequencing generated sequences of the most abundant organisms and varied by primers. Discolored moth (Renia discoloralis), varied carpet beetle (Anthrenus verbasci) and unidentified Gastropoda, Arthropoda, and Diploda sequences were discovered. Pilot sample dissections prior to molecular work allow higher order taxonomic assessment but no individual species were identified. We seek to resolve if NGS, which provides exponentially more data than cloning is the best means of informing diet. We further seek to quantify NGS detection rates for Lepidoptera, a widely reported high value food for breeding birds, and other soft-bodied prey that are rarely detected by scat dissections; and to test if visually confirmed present prey are detected. Our data help to validate a low impact molecular method that will improve our understanding the wood thrush’s niche and which can be applied to studies of other bird species.