Two major genomic reference collections, Repbase and Dfam, are merging into a unified open-access framework to support transposable element identification and classification. For over three decades, genomic research spanning genome assembly, gene annotation, evolutionary analysis, and expanding biomedical and agricultural applications has depended on these complementary resources. Under the newly announced partnership, the Genetic Information Research Institute is releasing the complete Repbase collection under a Creative Commons Zero public domain license and integrating its database with Dfam.
Repbase was established in 1990 by Jerzy Jurka at the Linus Pauling Institute of Science and Medicine before moving to the Genetic Information Research Institute in 1994. It has served as the primary reference collection of eukaryotic transposable element consensus sequences. Arian Smit was an early contributor whose RepeatMasker software became the library’s primary pathway to the genomics community. The resource gained prominence through meticulous expert curation, with researchers manually refining transposable element family boundaries, structural classifications, and consensus sequence reconstructions across diverse eukaryotic lineages.
Dfam was founded in 2012 as an open-access alternative resource representing transposable element families using profile hidden Markov models built from seed alignments of genomic instances. Because Dfam preserves underlying instances alongside their specific genomic coordinates, the underlying evidence supporting each family model remains transparent and traceable. Maintained by Robert Hubley and Arian Smit at the Institute for Systems Biology and Travis Wheeler at the University of Arizona, Dfam has expanded to house models for nearly 3,000 species and currently serves as the default open library for RepeatMasker.
Despite their shared utility, both databases previously evolved along divergent operational trajectories. Repbase maintained deep expert manual curation but operated under a restrictive licensing framework, eventually implementing user subscription fees in 2019 following widespread cuts to National Institutes of Health funding for biological databases. Conversely, Dfam provided open scalability through partnerships with efforts like the Vertebrate Genomes Project and the Zoonomia project, but it absorbed large volumes of de novo computational libraries that lacked the intensive manual curation typical of Repbase entries.
The integration represents the realization of a collaborative roadmap initially outlined in a 2016 publication, which proposed building profile hidden Markov models based on the Repbase-derived RepeatMasker library. As part of the transition, the core Repbase team from the Genetic Information Research Institute will join the Dfam team. To catalyze the administrative and technical migration, the Genetic Information Research Institute has provided a gift to the University of Arizona to fund personnel and operational requirements needed to construct the shared infrastructure.
Executing the merger requires addressing several technical bottlenecks across both platforms. Software engineers and bioinformaticians must establish conversion protocols to translate curated Repbase consensus sequences into Dfam’s profile hidden Markov model seed alignment format. The team must also reconcile overlapping and redundant transposable element family annotations across both databases while migrating specialized curation workflows from the Genetic Information Research Institute into Dfam’s existing software framework.
The release strategy will proceed in phased stages across future database versions. An initial open release will distribute the Repbase collection in its current format under the CC-0 public domain license, followed by progressive technical incorporation into the Dfam platform. Beyond integrating legacy data, the combined research group plans to develop semi-automated curation methods to manage the rapidly increasing volume of community-contributed de novo transposable element libraries.
The collaborative team also plans to expand community outreach by hosting dedicated workshops, tutorials, and updated web interfaces that preserve traditional Repbase functional features within Dfam’s open architecture. Researchers indicate that uniting Repbase’s curation depth with Dfam’s transparent, scalable profile hidden Markov model framework establishes an open resource intended to assist genomic annotation at all scales, from individual genome projects to multi-species comparative initiatives.
The historical relationship between the two initiatives spans decades, highlighted by Arian Smit and Jerzy Jurka meeting at the 2006 Genomic Impact of Eukaryotic Transposable Elements conference at Asilomar, twenty years before this unified integration. Jurka actively participated in the broad collaboration that originally developed Dfam, establishing an enduring connection between GIRI’s curation expertise and Dfam’s traceable framework. Over decades of database development, GIRI researchers and community contributors systematically characterized numerous transposable element superfamilies across diverse eukaryotic lineages, establishing foundational knowledge and consensus reference sequences that guided eukaryotic genomics.
Dfam’s institutional network also extends to partnerships with the European Bioinformatics Institute alongside other major sequencing initiatives. Because accurate identification and classification of transposable elements and their remnants remain foundational to genome assembly, gene annotation, evolutionary analysis, and expanding agricultural and biomedical applications, unifying these complementary systems resolves long-standing data challenges. The research community has long required both the curation depth of Repbase and the open scalability of Dfam. By channeling decades of expert knowledge accumulated at GIRI into an unrestricted repository, the collaborative team is creating a single, coherent resource. Uniting Repbase’s manual curation depth with Dfam’s scalable infrastructure and broad taxonomic coverage is expected to assist researchers across every scale of investigation, whether they are annotating an individual newly sequenced genome or generating comprehensive repeat libraries across hundreds of species.
Subject of Research: Integration of Repbase and Dfam into an open-access transposable element database
Article Title: Unifying Repbase and Dfam: a new open foundation for transposable element research
Article References: Kojima, K. K., Smit, A. F. A., Bao, W., Kohany, O., Kojima, N. F., Jurka, T., Hubley, R., & Wheeler, T. J. (2026). Unifying Repbase and Dfam: a new open foundation for transposable element research. Mobile DNA, 17(1), Article 17. https://doi.org/10.1186/s13100-026-00409-9
Image Credits: AI Generated
DOI: 10.1186/s13100-026-00409-9
Keywords: transposable elements, Repbase, Dfam, RepeatMasker, genomics, genome annotation, bioinformatics, hidden Markov models, open access, Genetic Information Research Institute, Institute for Systems Biology, University of Arizona, consensus sequences, seed alignments, computational biology, eukaryotic genomes
Cite Scienmag News
Juliet Wilcox. (November 21, 2026). Repbase and Dfam Integrate to Create Unified Open Database for Transposable Elements. Scienmag. https://scienmag.com/repbase-and-dfam-integrate-to-create-unified-open-database-for-transposable-elements/
Juliet Wilcox. "Repbase and Dfam Integrate to Create Unified Open Database for Transposable Elements." Scienmag, 21 November 2026, https://scienmag.com/repbase-and-dfam-integrate-to-create-unified-open-database-for-transposable-elements/. Accessed 21 November 2026.
Juliet Wilcox. "Repbase and Dfam Integrate to Create Unified Open Database for Transposable Elements." Scienmag. November 21, 2026. https://scienmag.com/repbase-and-dfam-integrate-to-create-unified-open-database-for-transposable-elements/








