Meeting Abstract

P1-125    Sea Snacks From DNA Tracks: DNA Metabarcoding to Characterize Green Sea Turtle Diet Sarkis, CM*; Seney, E; Hoenig, BD; Gaspar, SA; Forsman, A; Department of Biology, University of Central Florida, Orlando, Florida, USA; Department of Biology, University of Central Florida, Orlando, Florida, USA; Marine Turtle Research Group, Department of Biology, University of Central Florida, Orlando, Florida, USA; Department of Biological Sciences, Duquesne University, Pittsburgh, PA, USA; Department of Biology, University of Central Florida, Orlando, Florida, USA; Department of Biology, University of Central Florida, Orlando, Florida, USA; dGenomics & Bioinformatics Cluster, University of Central Florida, Orlando, Florida, USA c.sarkis31@knights.ucf.edu

Sea turtle diet varies among species, life stages, and localities and includes both plants and animal prey items. Visual and microscopic examination of lavage, gut content, and fecal samples are commonly used for diet identification. However, visual ID of food items can be challenging due to physical degradation. DNA metabarcoding makes it possible to identify diet sample constituents based on genomic targets that are conserved across taxa but vary in nucleotide sequence. The objective of this study was to identify and optimize primer sets for efficient amplicon library preparation that maximizes the prey taxa detected through DNA metabarcoding for green sea turtles (Chelonia mydas). Samples consisted of 39 colon homogenates collected from stranded turtles (FL, USA). We tested six primer sets, targeting five genomic regions, previously developed to detect plants, animals, or eukaryotes more broadly: CO1, 18sV1-3, 18sV4, UPA, rbcL, and ITS. Efficiency of each primer set varied based on (a) amplification efficiency during PCR, (b) number of raw reads produced by Illumina sequencing, and (c) number of reads retained after quality control using Cutadapt and DADA2. During library prep, CO1 was most effective, with consistently strong amplification across samples, whereas rbcL required multiple reamplifications. After sequencing,18sV1-3 and rbcL had the greatest number of samples with < 1000 reads, thus, appearing less reliable. During quality filtering of reads, ITS and 18sV1-3 libraries lost substantial numbers of reads. Overall, the rbcL primer pair was difficult to work with at several steps in library prep and read processing, followed by 18sV1-3. The CO1 and 18sV4 primer sets, on the other hand, were both efficient at PCR, sequencing, and read processing steps. This presentation will also discuss the taxonomic classifications made by the different primer sets and the efficiency.