Domain-Level Annotation and Conservation of Human Endogenous Retroviruses in the T2T-CHM13 Genome

💡 Introduction This dataset provides a complete, genome-wide annotation of conserved retroviral protein domains within human endogenous retroviruses (HERVs) mapped to the telomere-to-telomere (T2T-CHM13) human reference genome. Using the same fully reproducible pipeline established for the GRCh38 version—based on RepeatMasker, EMBOSS getorf, HMMER, and InterProScan—we analyzed all open reading frames (ORFs) predicted within internal HERV regions annotated across the T2T assembly. Because T2T-CHM13 resolves previously inaccessible regions (including centromeres, telomeres, large segmental duplications, and alpha-satellite arrays), this dataset expands the catalog of internal HERV ORFs and their conserved domains beyond what was possible with GRCh38. We identified 17,946 retroviral-like domain hits, including domains belonging to reverse transcriptase, RNase H, integrase, protease, Gag, Env, and accessory genes. As in the GRCh38 analysis, thousands of these domains show moderate to high conservation, and many exhibit near-complete alignment to their reference HMM profiles. InterProScan and Phobius predictions provide additional information about catalytic residues, structural motifs, and transmembrane features, offering a detailed assessment of potential functional retention in T2T-specific or previously unresolved loci. 🔭 Integration into HERVarium This dataset constitutes the T2T internal-domain component of HERVarium, a comprehensive reference resource for Human Endogenous Retroviruses that integrates: internal region protein-domain annotations (this dataset), and LTR regulatory annotations (U3/R/U5 segmentation, motif content, PBS/PPT predictions). The LTR companion dataset (at the moment, only available for GRCh38) is available at:https://doi.org/10.5281/zenodo.17602210 Together with the GRCh38 version, the T2T dataset enables comparative analyses of HERV coding potential across human assemblies and provides coverage of regions previously unresolved or missing in GRCh38. 🧬 The dataset includes: BED files with genomic coordinates of all annotated domains FASTA files containing predicted ORFs and extracted domain sequences InterProScan outputs with domain and motif annotation Phobius results for ENV transmembrane and signal peptide prediction Summary tables describing domain conservation and co-occurrence architecture RepeatMasker output for the T2T assembly used in the analysis GyDB HMM profile database and classification table, required to reproduce the pipeline These resources provide a high-resolution map of the protein-coding legacy of HERVs on the most complete human genome available to date. 📄 Description of each file HERV_t2t_orfs_aa_sequences.fasta: amino acid sequences of predicted ORFs within internal regions. HERV_t2t_hmmscan_output.tbl: raw hmmscan output with detected GyDB domain matches. HERV_t2t_loci_annotated_domains.tsv: filtered table summarizing the best conserved domain per domain class per ORF (coverage, score, e-value). HERV_t2t_domains.bed: genomic coordinates of mapped domains on the T2T genome. HERV_t2t_domains_nt_sequences.fasta: nucleotide sequences of domain-mapped regions. HERV_t2t_domains_aa_sequences.fasta: amino acid sequences of domain regions from the original ORFs. HERV_t2t_interproscan_output.zip: full InterProScan annotation directory (XML, TSV, CDD, etc.). HERV_t2t_interproscan_cdd_parsed.tsv: parsed Conserved Domains Database summary. ENV_t2t phobius_output.txt: Phobius predictions for ENV-domain-containing ORFs. HERV_t2t_internal.bed: genomic coordinates of reconstructed internal regions. gydb_domains_classification.tsv: maps each HMM profile to GAG/POL/ENV/Accessory/Other. combined_gydb.hmm: concatenated HMM profile database used for domain detection. GCF_009914755.1_T2T-CHM13v2.0_genomic.fna.out.tar.xz: RepeatMasker output for the T2T assembly used in the analysis. (use your actual filename) 💻 Code and citation This dataset was generated using the same pipeline as the GRCh38 analysis, available at:🔗 https://github.com/funcgen/herv-domain-map If you use this dataset, please cite both the dataset DOI and the corresponding article: Montserrat-Ayuso, T., & Esteve-Codina, A. (2025). A comprehensive annotation of conserved protein domains in human endogenous retroviruses. bioRxiv. https://doi.org/10.1101/2025.07.25.666750 Version history v1: Initial release of internal-region annotations and domain-level conservation mapped to the T2T-CHM13 genome. Includes full domain annotation pipeline outputs, InterProScan and Phobius results, and the RepeatMasker output used in the analysis.