Fig. 3. Overall composition of Biosynthetic Gene Clusters (BGCs) of S. anthophorae sp. nov., S. nidicola sp. nov. and closest relatives. (a) Principal Component Analysis (PCA) plot of BGC composition, showing distinction of BGC composition based both on species and on subclade. Analysis was limited to differentially present BGCs. Colour represents actual or proposed species, labels indicate clade within S. anthophorae sp. nov. BH034^T and BH097 are identical in BGC composition and thus the points are directly overlapping. PCA made with prcomp() in R ‘stats’ package (scale=T). (b) Venn diagram showing total differentiating and shared BGCs between bee isolates and close relatives (S. anthophorae sp. nov., two clades, S. nidicola sp. nov., S. endophyticus and S. fractus). Total number of BGCs in each species or group indicated in parentheses under name. S. anthophorae sp. nov. as a species (clades 1 and 2) has the most unique BGCs at 13* (circled in green) when compared to S. endophyticus YIM 65594^T and S. fractus MV32^T. Clade 1 represents BH034T and BH097, clade 2 represents BH055, BH104 and BH105. Unique BGCs were not necessarily found in all strains within the clade. There are no BGCs that specifically distinguish all bee isolates from the other closest species (circled in yellow). *One of these 13 overlaps with a BGC found in S. kunmingensis DSM 41681^T (not shown), leaving 12 unique BGCs when compared to the three closest relatives (see Table S3). Made with InteractiVenn [68].