Fig. 5

F. prausnitzii phages in healthy versus IBD-diseased individuals from the dataset of Norman et al. [18]. a Workflow of the procedure. a To infer the presence of a phage in a given sample, virome reads were aligned to the phage genome, and hits showing greater than 75% identity were retained. b If hits of a given sample were concentrated on less than 5% of the phage genome, the sample was not considered further for this phage. c The number of hits was compared to the maximal theoretical number of hits that could result from the measured bacterial DNA contamination. If the number of hits was significantly higher, the sample was classified as phage-positive. d If phage-positive samples represent less than 3% of total samples, this phage was not further studied. e The proportion of phage-positive viromes is compared to IBD versus healthy viromes. f In phage-positive viromes, phage frequency is compared to IBD versus healthy samples. b Fraction of phage-positive samples according to the health status of individuals. Three phages are significantly more prevalent in IBD-diseased than in healthy individuals. The highly prevalent crAssphage was included as a positive control in the analysis. c Phage abundance (read frequency) in positive samples according to the health status of individuals. Two phages, Epona and LughL2/6, are significantly more abundant in IBD patients than in healthy individuals