How low can you go? Establishing detection limits for rare eukaryotes in Southern Ocean sedimentary ancient DNA

Sedimentary ancient DNA (sedaDNA) is genetic material extracted from paleoarchives. It provides insights into the composition and dynamics of ecosystems over time. Such information can be crucial in anticipating how ecological communities may respond to environmental shifts within the context of the current climate crisis. However, challenges exist in accurately verifying ancient DNA from ecologically significant vertebrate species (e.g. fishes, aquatic birds, and mammals). These species occur only in trace amounts in sedimentary records. Here, we benchmark a stringent bioinformatic pipeline using synthetic and empirical metagenomic sedaDNA data from IODP Expedition 382 (Scotia Sea). Our objectives are threefold: (i) test taxonomic assignment precision for rare marine eukaryotes, (ii) evaluate taxonomic assignment sensitivity across different sediment ages, and (iii) establish the minimum sequence quantity necessary for robust identification.

We demonstrate that taxonomic assignment precision varied significantly with sequence quantity and metagenomic context. Assignment sensitivity decreased with taxonomic rank and database representation. Reliable detection of low-abundance taxa in sedaDNA is achievable with 250 and 500 DNA fragments at the family and genus level, respectively. The reanalysis of IODP Exp. 382 sedaDNA data, using a custom built marine vertebrate-focused reference database, resulted in the first genetic reconstruction of the vertebrate community in the Scotia Sea. This lays the groundwork for future investigations into the presence and biodiversity of Southern Ocean vertebrates using sedaDNA.

All project related scripts and generated simulated datasets are available in ae_fishing_benchmark repository (https://github.com/33davis/ae_fishing_benchmark). The demultiplexed raw data in relation to the IODP Exp. 382 U1538 reanalysed during this study is available in the NCBI Sequence Read Archive database (https://www.ncbi.nlm.nih.gov/sra) under Accession code/BioProject PRJNA861836 (BioSamples SAMN29928044 - SAMN29928123) and includes metadata for each sediment and control sample.