Gene deletion block size distribution. (a) Frequency of gene deletion blocks with a given number of contiguous absent genes (%) for 11 gut bacterial species. Each point corresponds to the mean frequency across 10 individuals for a given species. (b) Length of different large deletion blocks across gut bacterial species. Several large deletion blocks are associated with prophage and conjugative elements and one containing PUL and CPS loci. (c) Heatmap showing an example of a large deletion block (containing 73 genes) in Bacteroides thetaiotaomicron species across the 41 individuals and four sequenced reference genomes. Metagenomic samples are represented in blue and reference genomes are represented in purple. The genes are represented by its NCBI sequence identifier (GI). The labeling of the reference genome used for metagenomic samples mapping is highlighted in bold. SusC/SusD and CPS annotations are based on those described in Xu et al. . Upstream of the CPS there are three SusC/SusD genes, which can be associated with one or more PULs, and downstream of the CPS there is another PUL. These PULs are associated with plant carbohydrate degradation based on eggNOG and MetaCyc annotation (cf ). For CPS and both PULs related sub-regions, where the loci are present, they show a conserved modularity in all individuals except for one individual (where one SusC/SusD was missing). In agreement with metagenomic samples in the completely sequenced genomes, CPS and PULs loci are present in three genomes and absent in another one.