A crystallographic fragment screening (CFS) has been performed on a spliceosomal yeast protein-protein complex of Aar2 and the RNaseH-like domain of Prp8 (AR). The F2X-Universal Library is a fragment library representing the commercially available chemical space of fragments. 917 fragments have been individually screened via crystal soaking. The datasets that could be successfully auto-processed and auto-refined had been subjected to a Pan-Dataset Density Analysis (PanDDA) (Pearce et al., 2017). This analysis allows to find low occupancy binders. The data has been analyzed in different ways; once all datasets were given as input to PanDDA and once the data was clustered via cluster4x (Ginn, 2020) and the individual clusters were given as input for PanDDA. After the analysis, in total 269 hits could be identified in the PanDDA event maps. Most fragments cluster in certain regions on the protein surface, which are termed binding sites. Ten binding sites were identified. Some of these binding sites overlap with known protein-protein interactions sites, while others have no known function. These novel binding sites could be potential interaction sites too. Furthermore, due to the repeated binding of individual fragment hits in the same binding site, certain structural overlaps could be observed of fragments. These provide the confirmation of binding modes which offers additional information for further compound development. The data provided here includes all folders of the different PanDDA-runs in individual tar.gz files. For each run the input (auto-refined structures, fragment structure files) and output data (pandda models and event and Z-maps) is given. Additionally, a directory overview as a PDF is provided to help navigate through the data. With this data, every identified fragment hit can be inspected individually.