Biodiversity and DNA degradation patterns recorded in a 2200-year-long sequence of sedimentary ancient DNA from the afrotropical Bwindi Impenetrable Forest, Uganda

Most of the Earth’s biodiversity is concentrated in the tropics. While the ultimate causes of this geographic pattern remain to be established, ongoing anthropogenic impacts in the tropical belt lead to rapid losses of species diversity. Ancient DNA approaches may help in deciphering temporal patterns in the diversification of tropical biota and could potentially provide historical baseline data on the diversity and distribution of species in anthropogenically modified landscapes. However, studies of sedimentary ancient DNA (sedaDNA) are thus far extremely rare in tropical settings and consequently its value as a conservation tool for tropical ecosystems remains to be tested systematically. To address this issue we present meta-genomic records of shot-gun sequenced sedimentary ancient DNA (sedaDNA) from several sediment cores from the equatorial Bwindi-Impenetrable Forest in Uganda. Because Bwindi is one the most diverse rainforests in Africa and its biota is well documented (including endangered species such Mountain Gorilla and Chimpanzee) it is well suited for a baseline study. We describe the taxonomic composition of sedaDNA from Bwindi for the past 2200 years at an average resolution of 50 years – one of the first comprehensive sedaDNA records of plant and animal taxa from a tropical rainforest. We specifically address the following questions: 1) How precisely can the taxonomic level of shotgun-sequenced tropical sediments be resolved at present? 2) What is the effect of temperature, acidity, nutrient availability, elemental and lithological sediment composition, and burial age on the degradation of DNA? Taxonomic assignments are based on three metagenomic classifiers and four reference databases and their reliability tested against local pollen and modern animal occurrence data. We find that 92.3% of our metagenomic data is taxonomically not identifiable due to the substantial underrepresentation of tropical taxa in genomic reference databases. Yet at ordinal level we reconstruct typical afrotropical assemblages, which do not decline in diversity over time. Our comprehensive set of ecological and sedimentological parameters including sediment age, surface water chemistry, pH, soil temperature, sediment density, sediment water and organic matter content, XRF elemental chemistry, nutrient concentrations, and magnetic susceptibility reveals that DNA degradation cannot be explained by any sedimentary parameter alone, is at Bwindi independent of sediment type, and most likely primarily driven by burial age, suggesting that DNA taphonomic models need to be site-specific in tropical environments. The viability of sedaDNA as a conservation-biology tool requires comprehensive genomic surveys of tropical biota to drastically improve the taxonomic representativeness of DNA reference databases.