Analysis of environmental DNA detected in lake sediments shows promise to become a great paleoecological technique that can provide detailed information about organism communities living in past environments. However, when interpreting sedimentary environmental DNA records, it is of crucial importance to separate ecosystem responses to large-scale environmental change from ‘noise’ caused by changes in sediment provenance or potential post-depositional DNA mobility. In this study, we show that plant and mammalian communities reconstructed from sediments are strongly affected by sediment provenance, but unaffected by vertical mobility of DNA after sediment deposition. We observe that DNA from aquatic plants was abundant in background sediment, while embedded detrital event layers (sediment deposited under erosion events) primarily contained terrestrial plants; hence, vertical mobility of aquatic plant DNA across sediment layers was negligible within our studied lakes. About 33% of the identified terrestrial plant genera were only found in detrital sediment suggesting that sediment origin had a strong impact on the reconstructed plant community. Similarly, DNA of some mammalian taxa (Capra hircus, Ursus arctos, Lepus, Felis) were only or preferentially found in detrital event layers. Temporal changes across the Holocene were the main drivers of change for reconstructed plant communities, but sediment type was the second most important factor of variance. Our results highlight that erosion and sediment provenance need to be considered when reconstructing past mammalian and plant communities using environmental DNA from lake sediments.

The uploaded material contains sediment DNA metabarcoding data targeting plants (trnL gene) and mammals (16S gene). It consists of two data sets:

1. Plant and mammalian DNA sequences from a Holocene sediment sequence that was retrieved from Lake Grosssee, Flumserberg, Eastern Switzerland.
2. Plant and mammalian DNA sequences from surface sediments that were recovered from seven lakes across the Central Alps.

The samples in each data set are paired to contain a sample of hemipelagic background sediment and a detrital event layer deposited during a heavy precipitation event, which is of similar age compared to the background sediment.

All DNA libraries include extraction blanks (named ‘K1_date’ and ‘K2_date’ for each of the extraction runs, respectively), which were run along the extractions with no sediment input, as well as PCR negatives (named ‘Neg1-8’), which contain PCR mix and primers, but no DNA. Sample tags are dual-unique (twin tagging) and hence, PCR blanks are not specifically outlined in the ngs files (that is, unused tag combinations in a library). More information on sample labelling and pairing can be found in the file ‘SampleInfo.xlsx’.

Coring and sampling

A sediment sequence spanning the Holocene was retrieved from Lake Grosssee, Flumserberg, Eastern Switzerland, with a percussion piston-coring system (Uwitec, Austria) in 2017. Five 3-m-long core segments were recovered with a horizontal offset of 1 m and by using a vertical offset by 0.5 m down to 8 m sediment depth. The established composite section uses the highest-quality sediments from these core segments and spans 7.73 m. Cores were split lengthwise and kept in a fridge until subsampling for sediment DNA analysis in 2021. Dark, organic-rich sediments are interpreted to represent hemi-pelagic background sedimentation, while reddish siliciclastic beds are interpreted as flood event deposits (Glaus, 2018). In 2016, a strong summer thunderstorm at Lake Grosssee triggered high runoff in all streams entering the lake as well as overland flow across the alpine meadows, leading to the deposition of a 2.5-cm-thick layer in the lake. The Holocene record contains more than 100 comparable event layers ranging from 0.2-10 cm in thickness, of which the youngest could be correlated to documented flood events in the region (Glaus, 2018). Nine of these event layers together with adjacent background sediments were sampled for eDNA analysis. Average age difference in the sample pairs is 102 years, based on an age-depth model, which was built on 17 radiocarbon dates from terrestrial plant macrofossils (Dwileski, 2022). The sample pairs span an age range from modern to around 8,000 calibrated years before present (cal. yrs BP).

Surface sediments were taken with a gravity corer (Uwitec, Austria) from the lake’s depocentre in seven lakes across the Central Alps in 2020. Cores were kept in a fridge until splitting and subsampling for sediment DNA analysis in 2021. Event layers were identified visually, based on characteristics such as fining-upward sequences, high minerogenic content in contrast to organic-rich surrounding sediment, and the local abundance of plant macrofossils. Sample pairs were spaced 1.5 cm apart, on average.

DNA extraction and sequencing

Sample pairs for sediment DNA analysis were taken using sterile sampling equipment in a lab where no polymerase chain reactions (PCR) are performed. We cleaned the core surfaces before subsampling and avoided sediment touching with the core liner. Samples of background sediments span across 2 cm, while event samples are an integration of the respective event layer (spanning 1-10 cm).

DNA from 0.5 g of wet sediment was extracted in a dedicated clean lab for ancient DNA analysis with a Qiagen PowerSoil Pro kit, following the manufacturer’s instructions, but incubating the samples on a shaker overnight at 60°C (lysis step). Two extractions were done per sediment depth and two negative extraction controls were added to each extraction batch run to detect potential contamination during extraction. DNA from all extracts, including extraction controls, was amplified in 4 PCR replicates, yielding 8 replicates per sediment depth. PCR reagents are summarised in Table 1. Plant DNA was amplified with primers targeting the P6 loop region of the chloroplast trnL gene (g, 5’-GGGCAATCCTGAGCCAA-3’; and h, 5’-CCATTGAGTCTCTGCACCTATC-3’; Taberlet et al. (2006)). The PCR protocol included 45 cycles of a denaturing step (95°C, 30 s), an annealing temperature of 55°C (30 s), and an elongation step at 72°C (30 s). Mammalian DNA was amplified with primers targeting the mitochondrial 16S gene (MamP007F, 5’-CGAGAAGACCCTATGGAGCT-3’; MamP007R, 5’-CCGAGGTCRCCCCAACC-3’; Giguet-Covex et al. (2014)) with a human blocker (MamP007_B_Hum1, 5’-GGAGCTTTAATTTATTAATGCAAACAGTACC-C3spacer 3’). The PCR protocol included 45 cycles (as suggested by Giguet-Covex et al. (2014)) of a denaturing step (95°C, 30 s), an annealing temperature of 50°C (30 s), and an elongation step at 72°C (30 s). Primers were twin-tagged on the 5’ end using a unique combination of 12 bp long nucleotide sequence with at least three mismatches (Binladen et al., 2007), plus 4 additional random nucleotides. The twin-tagging approach was applied to avoid effects of tag jumping, which can greatly distort eDNA interpretations (Rodriguez-Martinez et al., 2023). On each PCR 96-well plate, we randomly arranged 7-8 PCR negatives to trace possible cross-contamination during PCR. After PCR, samples, PCR negatives, and extraction blanks were pooled by plate (equivolume) and purified with a QIAQuick PCR purification kit (Qiagen, Germany). Whenever possible, sediment sample pairs were extracted on the same day and arranged in the same library to avoid any bias through downstream analyses. Library preparation was done by Novogene (UK) following the NEBNext Ultra II DNA Library Prep Kit for Illumina (NEB #E7645) without the PCR enrichment step and samples were sequenced on an Illumina NovaSeq 6000 PE150 platform. Libraries with the same primer tags were sequenced on separate sequencing lanes to avoid the effects of index hopping.

Table 1 Reagents used for PCR reactions to amplify plant and mammalian DNA, respectively.

References

Binladen, J., Gilbert, M. T., Bollback, J. P., Panitz, F., Bendixen, C., Nielsen, R., and Willerslev, E., 2007, The use of coded PCR primers enables high-throughput sequencing of multiple homolog amplification products by 454 parallel sequencing: PLoS One, v. 2, no. 2, p. e197.

Dwileski, A., 2022, Holocene vegetation change in the eastern Swiss Alps (Grosssee, SG): Effects of climate and human impact [Master of Sciences: University of Basel.

Giguet-Covex, C., Pansu, J., Arnaud, F., Rey, P. J., Griggo, C., Gielly, L., Domaizon, I., Coissac, E., David, F., Choler, P., Poulenard, J., and Taberlet, P., 2014, Long livestock farming history and human landscape shaping revealed by lake sediment DNA: Nat Commun, v. 5, p. 3211.

Glaus, N., 2018, Holocene flood events and sediment flux in Alpine lake Grosssee (Flumserberg, Switzerland) [Master of Science: University of Bern, Switzerland.

Rodriguez-Martinez, S., Klaminder, J., Morlock, M. A., Dalen, L., and Huang, D. Y., 2023, The topological nature of tag jumping in environmental DNA metabarcoding studies: Mol Ecol Resour, v. 23, no. 3, p. 621-631.

Taberlet, P., Coissac, E., Pompanon, F., Gielly, L., Miquel, C., Valentini, A., Vermat, T., Corthier, G., Brochmann, C., and Willerslev, E., 2006, Power and limitations of the chloroplast trn L (UAA) intron for plant DNA barcoding: Nucleic Acids Research, v. 35, no. 3, p. e14-e14.

Raw sequencing data in .fq.gz format. Files need to be decompressed and can be run through an Obitools3 pipeline, for example.