Kmer approaches have greatly developed in recent years, largely driven by the advent of next-generation sequencing. Their speed and automatism, since they are annotation independent, are major advantages. The interest of applying k-mer approaches to study the mobilome is only starting to be explored, with many efforts oriented towards viral metagenomics [1, 2]. We evaluated the potential of applying simple k-mer approaches to understand the evolutionary history of mobile elements by focusing on viruses and plasmids from the domain Archaea, which includes specific and well-defined extrachromosomal element families [3]. We selected more than 590 cell, virus and plasmid genomes, originating from 11 distinct orders of archaeal cells. We subsequently implemented multivariate and statistical analyses to identify the factors underlying the genome 5-mer composition of mobile elements in the domain Archaea. The mobile element families, the genome GC contents and the phylogenetic position of the host at the taxonomic level of the order were identified as major explanatory factors. Genomes overall grouped according to the host order, except for the haloarchaea, all belonging to the class Halobacteria, which formed a single supergroup. Within each group, cells tended to cluster together while viruses and plasmids tended to cluster according to their own taxonomic family. This was only a general trend as a number of exceptions were observed, with an inhomogeneous distribution across the dataset. The observed pattern likely results from the combined influence of co-evolution and environmental constraints, highly contrasted among those archaea which comprise methanogens and diverse extremophiles. It confirms the potential of k-mer signal for extrachromosomal contig analysis [1, 2, 4], in particular for their taxonomic assignation. Another application could be the detection of singular evolutionary trajectories by focusing on outliers. Indeed, previously-known and one new case of recent host transfers were efficiently detected during the present study. K-mer approaches rely on a distinct informational content than the more elaborate gene sharing networks and appear as complementary. References 1.Galiez C et al, Bioinformatics 2017, 33:3113-3114. 2.Iranzo J et al, Journal of Virology 2016, 90:11043-11055. 3.Ren J et al, Microbiome 2017, 5(1):69. 4.Krawczyk PS et al, Nucleic Acids Research 2018:gkx1321-gkx1321.