Comparative analysis of long identical DNA sequences shared by bacterial genomes belonging to different genera showed that horizontal gene transfer (HGT) across deep taxonomic boundaries — including across phyla — is frequent and widespread, with a statistical signature (a power-law match-length distribution with exponent −3-3) from which an 'effective HGT rate' can be estimated between any pair of taxa. This effective rate, however, is a lumped parameter that conflates the physical transfer process, the length of transferred segments, the probability of establishment in the recipient population, and mutation rates, and it varies by orders of magnitude across taxon pairs and gene functions for reasons that remain unexplained. The open problem is the mechanistic and causal account of long-distance HGT: which molecular mechanisms (mobile genetic elements, phages, natural transformation, or other routes) mediate the exchange of genetic material between evolutionarily distant bacteria, and what the independent, causal contributions of ecological co-occurrence, genomic compatibility (GC content, Gram staining), gene function, and post-transfer establishment probability — distinguished from the physical transfer rate itself — are to the observed variation in transfer rates.

Background

Horizontal gene transfer (HGT) is the acquisition of genetic material by means other than parent-to-offspring inheritance — in bacteria mainly via conjugation (transfer through direct cell-to-cell contact, usually mediated by plasmids), transduction (packaging and transfer of DNA by bacteriophages), and natural transformation (uptake of extracellular DNA). HGT is a dominant force in bacterial genome evolution, most conspicuously in the spread of antibiotic resistance genes. While transfer between close relatives is well characterized, a global, quantitative picture of exchange between highly divergent bacteria was long missing. Sheinman, Arkhipova, Arndt, Dutilh, Hermsen and Massip (eLife 2021; bioRxiv preprint 2020) filled this gap by scanning 93,481 bacterial genomes from the NCBI RefSeq database (the curated reference-genome collection of the U.S. National Library of Medicine) for maximal exact DNA matches (identical stretches found via alignment-free comparison, longer than 300 bp) between organisms of different genera. Because transferred DNA accumulates mutations after transfer, such long exact matches flag recent events, within an estimated detection horizon of roughly 1000 years. The authors identified 138,273 HGT events; about one third of organisms (32.6%) showed evidence of exchange across family boundaries, and about 8% of organisms across phylum boundaries, within this recent window. Statistically, the match length distribution (MLD — the number of shared identical stretches as a function of their length) follows a robust power law with exponent −3-3 over many decades of match length. A mathematical model explains this exponent under minimal assumptions, and its prefactor yields an 'effective HGT rate' between any taxon pair. Crucially, this rate is a lumped parameter combining the physical transfer rate per genome length, the typical transferred sequence length, the degree to which a transferred sequence becomes established (fixed or maintained) in the populations of both taxa, and the effective mutation rate; the model therefore cannot separate how often transfer happens from how often transferred DNA persists. Three observations demanded explanation and were explicitly left open by the source. First, HGT rates vary dramatically between taxon pairs, and correlate simultaneously with ecological environment, GC content (the genomic proportion of G/C base pairs, a proxy for cellular compatibility differences), and Gram staining (a bacterial cell-wall classification that determines, among other things, which mobile elements can adsorb or conjugate); because these properties are correlated with each other, their independent causal contributions could not be disentangled from the data. Second, the molecular mechanisms that physically enable DNA to cross phylum boundaries — the mobile genetic elements (plasmids, bacteriophages, transposons and other DNA molecules that move within or between genomes) or transfer vectors involved — remain unknown; the source proposed dedicated studies of families with very high exchange rates as a route in. Third, the model's key structural assumption is that transfer occurs continuously at a non-zero rate up to the present, while episodic 'bursts' of transfer or gene-loss mechanisms could also explain deviations from the −3-3 law. Follow-up work has since clarified parts of this picture: the same group showed theoretically and empirically that episodic transfer would produce exponential MLD tails while sustained continuous transfer over a network with high-transferability hubs produces the observed power laws (Genome Biol Evol 2026), and extended modeling beyond the 1000-year horizon (PNAS 2024); large-scale analyses established that co-occurring, interacting, high-abundance species exchange more genes (Nat Ecol Evol 2024) and that inter-phylum acquisitions are mostly lost quickly, with a functionally biased subset persisting long-term (Mol Biol Evol 2026); and a systematic survey of inter-phylum transfers of antibiotic resistance genes found that conjugative systems are seldom shared across phyla, implying that other, still-unidentified mechanisms dominate long-distance dissemination (mSphere 2025). What no study has delivered is the mechanistic identification of the long-distance transfer routes or a causal decomposition of the determinants of inter-taxon HGT-rate variation — the questions the source itself posed.

Problem Statement

For horizontal gene transfer between evolutionarily distant bacteria — in particular transfers crossing phylum boundaries, as detected genome-wide through long identical DNA sequences shared by divergent taxa — determine (1) the molecular mechanisms and mobile genetic elements that mediate such long-distance transfer: which vectors or transfer routes physically carry DNA across deep phylogenetic barriers; and (2) the causal decomposition of the observed orders-of-magnitude variation in inter-taxon HGT rates: the independent contributions of ecological co-occurrence, genomic compatibility covariates such as GC content and Gram staining, gene function, and post-transfer establishment probability, distinguished from the physical transfer rate itself rather than absorbed into a single lumped effective-rate parameter. A resolution of (1) is a named transfer vector or route — a specific mobile genetic element, phage, or transformation mechanism — with evidence tying it to the observed inter-phylum transfers, or an explicit finding that no shared vector accounts for them; a resolution of (2) is a decomposition that reports the independent contribution of each named determinant (ecological co-occurrence, GC content, Gram staining, gene function, and post-transfer establishment probability) and separates post-transfer establishment and mutation from the physical transfer rate, rather than an aggregate of the correlated covariates.

References

  1. Sheinman M, Arkhipova K, Arndt PF, Dutilh BE, Hermsen R, Massip F. Identical sequences found in distant genomes reveal frequent horizontal transfer across the bacterial domain. eLife 2021;10:e62719. DOI: 10.7554/eLife.62719. https://doi.org/10.7554/eLife.62719
  2. Sheinman M, Arkhipova K, Arndt PF, Dutilh BE, Hermsen R, Massip F. Long identical sequences found in multiple bacterial genomes reveal frequent and widespread exchange of genetic material between distant species. bioRxiv 2020. DOI: 10.1101/2020.06.09.139501. https://doi.org/10.1101/2020.06.09.139501
  3. Sheinman M, Arndt PF, Massip F. Modeling the mosaic structure of bacterial genomes to infer their evolutionary history. PNAS 2024;121(13). DOI: 10.1073/pnas.2313367121. https://doi.org/10.1073/pnas.2313367121
  4. Sheinman M, Stentella T, Etheimer P, Massip F, Arndt PF. Reconstructing the network of horizontal gene exchange in bacteria to differentiate direct and indirect transfers. Genome Biology and Evolution 2026;18(5). DOI: 10.1093/gbe/evag099. https://doi.org/10.1093/gbe/evag099
  5. Dmitrijeva M, Tackmann J, Matias Rodrigues JF, Huerta-Cepas J, Coelho LP, von Mering C. A global survey of prokaryotic genomes reveals the eco-evolutionary pressures driving horizontal gene transfer. Nature Ecology & Evolution 2024;8:986-998. DOI: 10.1038/s41559-024-02357-0. https://doi.org/10.1038/s41559-024-02357-0
  6. Mishra S, Weit K, Lercher MJ. The Fate of Horizontally Acquired Genes: Rapid Initial Turnover Followed by Long-Term Persistence. Molecular Biology and Evolution 2026;43(7). DOI: 10.1093/molbev/msag146. https://doi.org/10.1093/molbev/msag146
  7. Parras-Moltó M, Lund D, Ebmeyer S, Larsson DGJ, Johnning A, Kristiansson E. The transfer of antibiotic resistance genes between evolutionarily distant bacteria. mSphere 2025;10(6). DOI: 10.1128/msphere.00114-25. https://doi.org/10.1128/msphere.00114-25
  8. Good BH, Bhatt AS, McDonald MJ. Unraveling the tempo and mode of horizontal gene transfer in bacteria. Trends in Microbiology 2025;33:853-865. DOI: 10.1016/j.tim.2025.03.009. https://doi.org/10.1016/j.tim.2025.03.009
  9. Van Etten J, Johnson MD. The ecology of horizontal gene transfer. Trends in Genetics 2026;42:539-554. DOI: 10.1016/j.tig.2026.02.002. https://doi.org/10.1016/j.tig.2026.02.002