Component of bacterial genomes are separated into the stable, relatively slow-evolving and species-specific chromosomes, and the accessory, exchangeable and rapid-evolving plasmids. The evolution of chromosomes is related to that of plasmids through their interaction into the genome. In recent years, phylogenetically diverse bacterial species have been shown to harbour multipartite stable genomes constituted of a primary chromosome, homologous to the standard bacterial chromosome (e.g., that of Escherichia coli), supplemented with one or more essential replicons adapted to the cell cycle. Although a plasmidic origin is most commonly postulated, these elements generally exhibit both chromosomal and plasmidic features. We hypothesized that stable additional replicons (megaplasmids and secondary chromosomes) represent evolutionary intermediates between chromosome and plasmid. We investigated how the different types of bacterial replicons might be related. Using all bacterial replicons for which a complete sequence is available, we focused on replication, partition and segregation since these functions are essential to the maintenance of the replicons, and are key to the emergence of multipartite genomes. We applied a global comparative genomic analysis of all bacterial replicons based on machine learning and data mining approaches to characterize the networks of clusters of homologous proteins implied in those processes. We thus bring to light characteristic replicon adaptations among the bacterial taxa known to harbour multipartite genomes (e.g., Vibrionaceae, Burkholderiaceae, Rhizobiales...) as well as in taxa such as Bacteroidetes and Firmicutes, in which additional replicons are scarcely represented. We further highlight global trends for the secondary replicons, and conclude on the singular status of these genomic elements.