Fig. 1: Identification and validation of tissue-resident microbiota in CRC patients.

a Schematic representation of the microbiome study design of 937 CRC patients from the Uppsala-Umeå Comprehensive Cancer Consortium (U-CAN) cohort. Patients included at Uppsala (UU, n = 775) and Umeå (UM, n = 162) University Hospitals served as discovery and internal validation cohort, respectively. Patients with primary colon cancer from the AC-ICAM (Atlas and Compass of Immune-Cancer-Microbiome Interactions in Colon Cancer) study (n = 246) served as external validation cohort. Tissue-resident microbes were identified using non-human reads from whole-genome sequencing (WGS) data, and were associated with host genomic alterations, gene expressions and patient survivals. Illustration elements (colon schematic and human outline silhouette) adapted from Servier Medical Art (https://smart.servier.com), CC BY 3.0. b Number of Kraken2-mapped prokaryotic reads from tumor tissues and paired blood samples (n = 462 pairs). c The proportions and taxonomic assignments of microbial reads classified as contaminants in tissue and paired blood (n = 462) samples from UU. d Number of annotated reads (left) and average genome coverage (right) of tissue-resident species across tumors (n = 775), normal adjacent tissues (NATs; n = 313) and blood (n = 462) samples from the UU cohort. e Average genome coverage (Y-axis) and number of mapped reads (X-axis) of tissue-resident prokaryotic species in tissue (n = 1088) and blood samples from the UU cohort. Each dot represents a different species. The name of the top 10 species by average Kraken2-mapped reads across tissue samples are shown. f Number of non-human reads from tumor tissue samples of the UU (n = 775), UM (n = 162) and AC-ICAM (n = 246) cohorts in the different processing steps. g Left: The average genome coverage of tissue-resident species in tumor tissue samples from the UU (n = 775), UM (n = 162) and AC-ICAM (n = 246) cohorts. Right: The top 50 most-covered species identified across three tumor datasets. In b, d, f, g, p-values were calculated using two-sided Wilcoxon rank-sum test. In all boxplots, boxes represent the interquartile ranges (IQRs) between the first and third quartiles, the center line represents the median, and the whiskers extend 1.5 times the IQR from the top and bottom of the box.