Global warming is causing irreversible changes to marine ecosystems, especially in the Arctic, where warming occurs up to three times faster than the global average. Therefore, there is a strong need to make accurate predictions about marine ecosystem responses to climate warming. To improve our predictions, we need to better understand the mechanisms that cause ecosystem turnovers, which can be achieved by looking at potential analogues in the past. In this project, I use sedimentary ancient DNA to study Middle and Late Pleistocene marine ecosystem turnovers and glacial-to-interglacial boundaries. I have generated shotgun metagenomics libraries for sediments from the Barents Sea, the Central Arctic Ocean, and the North Sea and I am planning to reconstruct whole marine ecosystems over the past 500 000 years, from primary producers, to zooplankton, fish, and even marine mammals. I will use these ecosystem reconstructions to compare ecosystem turnover at multiple glacial - interglacial transitions with the aim to test whether multiple phases of rapid warming during the Pleistocene had similar effects on marine ecosystems.