Fig. 2

The normalized number of bacterial read pair counts for each sample is represented in this heat map for all bacterial taxa with >20 reads per one million total reads. The dendrogram represents the agglomerative hierarchical clustering of the samples based on their bacterial reads. Color-coded bars are present to represent the cancer type, tissue source site, sequencing type, sample type, and sequencing center. GBM and OV samples have the lowest abundance of bacterial taxa, while the only samples showing increased read counts of Mycobacterium spp. are OV samples. Enterobacteriaceae and “other” were the dominant taxa in most cancer types, except for Ralstonia spp. in GBM, Ralstonia spp. and “other” in AML, and Enterobacteriaceae and Mycobacterium spp. in OV