Phylogenetic trees were inferred on all samples, where an appropriate reference sequence could be chosen. These trees present a higher level view, with one tree potentially containing samples belonging to different sub-types of the species. Therefore, the reference sequence name does not necessarily agree with the sub-type of the sample. Due to the distance of the reference sequences from the samples (max. 98 ANI%), they are not suitable for detecting outbreaks, however, it is expected that epidemiologically connected samples do cluster tightly together. The bioinformatics method is described in Szarvas et al 2021.