Fig. 5

Results for the principal component analysis and hierarchical clustering. They were realized with the variables: HT quantities and proportions in the studied genomes and their partners. The individuals represented are the 60 bacteria where HT were identified, the colors of points and numbers indicate their habitat. Two clusters were defined (transparent color squares) according the bacterial habitats (Chi2 test = 2.4*10-7): 91.3% of bacteria from vertebrates are gathered in the cluster 1 (blue square) and 92.0% of environmental bacteria belong to the cluster 2 (red square). Axis 1 contains 33.9% of the information and mainly represents the transfer partners (Variable represented: Cyanobacteria, Firmicutes, Fusobacteria and Alphaproteobacteria (>40%), Actinobacteria, Acidobacteria and Epsilonbacteria (>20%)). The axis 2 contains 21.0% of total information and represents mainly the HT quantities and proportions (Variables represented: Sequences quantities and proportions (>70%), Actinobacteria and Gammaproteobacteria (>15%)). However, 46.1% of the information is missing, concerning mainly the Gammaproteobacteria and Acidobacteria