Whole-genome sequencing analysis of the colorectal cancer microbiome. (A) Schematic of experimental and computational whole-genome sequencing analysis workflow. (B) Hierarchical clustering of phylotype relative abundance measurements demonstrates that microbial composition of tumor/normal pairs within individuals is more highly correlated than tumor/tumor pairs or normal/normal pairs from different individuals. (Green) Normal samples; (purple) tumors. (C) Linear discriminant analysis (LDA) coupled with effect size measurements identifies Fusobacterium as the most differentially abundant taxon in colon tumor versus normal specimens by whole-genome sequencing in nine individuals. Tumor-enriched taxa are indicated with a positive LDA score (black), and taxa enriched in normal tissue have a negative score (gray). Only taxa meeting an LDA significant threshold of 1.8 are shown. (D) Percent relative abundance for the genus Fusobacterium is depicted across all samples in the order of the labels in B, demonstrating a tumor-enrichment in most individuals.
