We performed a temperature incubation experiment on board the RV Polarstern with a unicellular microbial community sampled from the Hausgarten station IV in Fram Strait during the campagin PS126 on June 1st, 2021 (Soltwedel et al., 2021). The community was sampled with CTD-bound niskin bottles from a depth of 15 m (Hoppmann et al., under review) and, after filtering the seawater through a 150 µm net, incubated in triplicate on plankton wheels in three temperature-controlled containers for ten days. To mimick todays and potential future temperature conditions of the Arctic ocean, we chose a control temperature of 2 °C, an intermediate warming scenario of 6 °C, and an extreme warming scenario of 9 °C. The goal was to investigate the effects of concurrent warming and Atlantification and therefore we chose an Arctic-Atlantic mixed water mass as community origin. This dataset comprises the 18S- and 16S-metabarcoding data of the starting as well as the final communities. A total of 500 mL of sample water was carefully vacuum-filtrated (<-200 mbar) onto polycarbonate filters (0.8 µm nominal pore size, Nucleopore, Whatman, Maidstone, UK). Filters were put into 2 mL cryovials containing 650 µL of warm extraction buffer and stored at -80 °C. After cell disruption with a MagNa Lyser (Roche Diagnostics, Basel, Switzerland), DNA extraction was performed according to the manufacturer’s protocol using the NucleoSpin Soil extraction kit (Macherey-Nagel GmbH, Düren, Germany). DNA concentration was quantified using a NanoDrop 8000 spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA) and normalized to 5 ng µL-1. Amplicons of the variable region 4 (V4) of the 18S rRNA and 16S rRNA gene for eukaryotes and bacteria, respectively, were generated according to the standard protocol of amplicon library preparation (16S Metagenomic Sequencing Library Preparation, Part #15044223 Rev. B. Illumina, San Diego, CA, USA) using the forward primer CCAGCASCYGCGGTAATTCC and reverse primer ACTTTCGTTCTTGAT for 18S rRNA gene sequencing (Bradley et al., 2016) and the forward primer GTGCCAGCMGCCGCGGTAA and reverse primer GGACTACHVGGGTWTCTAAT for 16S rRNA gene sequencing (Fadeev et al., 2018), all including an Illumina tail. Single samples were indexed using the Nextera XT Index Kit v2 Set A primers (Illumina, San Diego, CA, USA), and the resulting libraries were pooled, one pool each for 18S and 16S rRNA gene sequencing. Sequencing was performed on a MiSeq sequencer (Illumina, San Diego, CA, USA), producing 300 base pair paired-end gene amplicon reads. Demultiplexing and FASTQ sequence file generation were carried out using the Generate FASTQ workflow of the MiSeq sequencer software. Primers were removed with v2.8 cutadapt (Martin, 2011), and further processing of the sequence data was performed using the v1.18 DADA2 R package (Callahan et al., 2016). In consideration of the read quality, which usually drops towards the 3'-end, the forward reads were trimmed after 240 to 260 base pairs, and the reverse reads were trimmed after 200 to 210 base pairs. For each pool, error rates were learned independently, and sequences were denoised. Paired-end reads were merged with a minimum overlap of 50 base pairs allowing 0 mismatches, and chimeras were predicted and removed. Taxonomic assignment of the resulting amplicon sequence variants (ASVs) was performed using the reference databases PR2 (v4.12.0) for eukaryotes (Guillou et al., 2013) and SILVA (v138) for bacteria (Quast et al., 2013). All scripts can be found on github (https://github.com/AntoniaAhme/PS126CommunityExperiment). The accompanying metadata are published on PANGAEA (https://doi.pangaea.de/10.1594/PANGAEA.960624).