Fig. 2

Heatmap of all taxa detected in all samples. Log2-transformed RPPB of all taxa in all samples (also including taxa that were filtered out subsequently as potential contaminants). Taxa were hierarchically clustered using Pearson correlation as a distance measure with average-linkage. Samples were hierarchically clustered within each project and type subgroup using Pearson correlation as a distance measure with average-linkage. Tumor tissue samples were all primary solid tumor biopsy material for all projects with the following exceptions: CLL-ES, where tumor tissue samples were blood-derived CLL cells; CMDI-UK, where 3 tumor tissue samples were bone-marrow-derived and 29 samples blood-derived; LAML-KR, where all 12 tumor tissue samples were bone-marrow-derived; PACA-AU, where 1 tumor tissue sample was cell-line-derived; and 96 solid tumor biopsy material and PRAD-UK, where 4 samples were cancerous lymph nodes, 1 sample was a metastatic lesion, and 28 samples were solid tumor biopsy material. Matched normal samples were all blood-derived for CLL-ES, LINC-JP, ORCA-IN, PRAD-CA, and RECA-EU, all matched non-cancerous tissue derived for BTCA-SG, LICA-FR, and PAEN-IT, a mix between blood-derived and matched non-cancerous tissue for BOCA-UK (69 vs. 7), BRCA-UK (44 vs. 1), ESAD-UK (87 vs. 10), LIRI-JP (250 vs. 6), PACA-AU (3 vs. 94), PACA-CA (55 vs. 68), and PAEN-AU (5 vs. 44), a mix between blood-derived and buccal cell-derived for CMDI-UK (31 vs. 1), a mix between blood-derived and EBV immortalized cell-line-derived for OV-AU (59 vs. 14), and a mix between blood-derived and cancer-free lymph node derived for PRAD-UK (23 vs.10). For some samples, the source tissue was not specified (BRCA-UK (n = 1), LAML-KR (n = 1), and PACA-CA (n = 24). All healthy donor samples are blood-derived