Fig. 6

Evaluation of clusters of contigs from the STEC outbreak microbiomes (see main text). (Left panel) Clusters’ content in pathogen (E. coli O104:H4) genome sequence. Peaks show the fraction of the E. coli O104:H4 genome embedded in each cluster computed by the CB-Cl module (k-mer size = 6, 300 output clusters (filtering out one dubious cluster containing more than 10 % of the original sequences); clusters are arbitrarily ordered on the x axis). (Right panel) Strength of the association between clusters and disease status. Cluster coordinates on the x axis are the same as in the left panel; peaks represent mutual information for each cluster with respect to the infection status of the microbiome donors (see main text). The cluster with the highest mutual information value encompasses about 70 % of the E. coli O104:H4 genome sequence