Structural Annotation and Transcription Factor Motif Maps for Human Endogenous Retroviruses

💡 Introduction This dataset provides a genome-wide, high-resolution annotation of long terminal repeats (LTRs) from human endogenous retroviruses (HERVs), including reconstructed U3–R–U5 segments, transcription factor binding motif (TFBM) maps, promoter features, and PBS/PPT flanking signals. Building on curated HERV internal annotations, we developed a reproducible pipeline to identify the structural and regulatory architecture of each LTR in the human reference genome (GRCh38). Starting from RepeatMasker annotations, we merged fragmented elements, classified each HERV-associated LTR as 5′, 3′, both, or solo, and delineated the U3, R, and U5 regions in a strand-aware manner. We then scanned each LTR sequence using FIMO and a JASPAR-downloaded TF motif library to quantify transcription factor binding site burden, identifying LTRs enriched for developmental- and proliferation-related regulators and depleted of neuronal-differentiation-associated TFBMs. Together, these annotations provide a detailed and unified view of the regulatory potential and structural organization of HERV LTRs across the genome. 🧬 The dataset includes: Genome-wide U3–R–U5 segmentation for all HERV-associated LTRs LTR regulatory maps with the TFBM found in each LTR PBS and PPT predictions based on tRNA-tail matching and purine-rich sequence detection Classification of each LTR as 5′, 3′, both, or solo, with links to internal regions BED files for all reconstructed segments and flanking regions These resources enable exploration of HERV LTR regulatory diversity, their potential co-option as promoters or enhancers, and their evolutionary relationship to host developmental programs. 📄 Description of included files HERV_LTR_fimo_results_tsv.tar.xz: tab-separated table of all transcription-factor motif hits detected by FIMO across annotated LTR sequences. Includes subfamily assignment, motif IDs (JASPAR 2024), genomic coordinates, match scores, p-values, q-values, and matched sequences. HERV_LTR_fimo_results_bed.tar.xz: BED file containing FIMO-detected transcription-factor binding sites mapped to the genome. Each entry reports genomic coordinates, motif identity, strand, and the FIMO match score in the BED score field. HERV_internal_LTR_map.tsv: table mapping each internal ERV region to its flanking 5′ and 3′ LTR components, including genomic coordinates, locus identifiers, LTR names, and distances, used to classify LTRs as 5′, 3′, tandem, or solo prior to U3/R/U5 annotation. HERV_LTR_U3_R_U5_segments_highconf.bed: BED file containing the genomic coordinates of each LTR’s annotated U3, R, and U5 segments. Each line represents one segment (U3, R, or U5) with the BED score field reporting the overall annotation confidence score derived from promoter and PAS evidence. HERV_LTR_U3_R_U5_segments_allconf.bed: BED file containing the genomic coordinates of all annotated U3, R, and U5 segments for every LTR, including both high-confidence and LOW_CONF cases. Each entry includes the segment type and confidence flag in the name field, and the BED score reports the annotation confidence derived from promoter and PAS motif evidence. HERV_LTR_U3_R_U5_catalogue.tsv: table summarizing U3/R/U5 structural annotations for all HERV-associated LTRs, including genomic coordinates, LTR role (5′, 3′, tandem, solo), linked internal regions, TSS and PAS positions, confidence scores, and detailed motif-based evidence used by the annotator. HERV_LTR_U3_R_U5_PBS_PPT.bed: BED file containing genomic coordinates of detected PBS and PPT elements flanking each LTR. The BED score column reports the length (in nucleotides) of the PBS or PPT motif. promoters_motifs.txt: curated list of promoter elements (TATA box, Inr, BRE, DPE, MTE, DCE, XCPE1, SP1), including consensus definitions and regular expressions used by the U3/R/U5 annotation script to score and identify transcription start sites in LTRs. hg38-mature-tRNAs.fa: FASTA file of all mature, intronless human tRNAs from GtRNAdb (including the CCA tail); used as the input to extract 3′ tRNA tails for building the PBS motif mini-library. tRNAs_manifest.tsv: table listing all mature tRNAs processed, their isotypes/anticodons, tail lengths, and output identifiers corresponding to the raw (non-deduplicated) tRNA tail fastas. tRNAs_dedup_manifest.tsv: manifest summarizing deduplicated tRNA tail sequences across RNA/DNA/RC spaces, including counts, representative headers, isotypes, anticodons, and tail statistics. tRNA_3prime_RNA.fa: 3′ terminal 20-nt tails (RNA alphabet) extracted from all mature GtRNAdb tRNAs; used as the source for PBS motif generation before conversion to DNA. tRNA_3prime_RNA_dedup.fa: deduplicated 3′ RNA tails from mature tRNAs, collapsing identical sequences across isotypes/anticodons; used to avoid redundant PBS matches. tRNA_3prime_DNA_revcomp.fa: deduplicated DNA 3′ tRNA tails, removing duplicates across isotypes; used by the PBS scanner to speed up and reduce redundant motif tests. tRNA_3prime_DNA_revcomp_dedup.fa: deduplicated reverse-complement DNA tRNA tails; the final PBS motif set used by 10_annotate_u3r_u5.py to detect primer binding sites downstream of 5′ LTRs. PAS_hexamers.txt: list of canonical and variant polyadenylation signal hexamers used by the U3/R/U5 annotator to detect PAS motifs and infer 3′-cleavage sites within LTRs. 🔭 Integration into HERVarium This dataset constitutes the LTR regulatory and structural component of HERVarium, a comprehensive reference resource for Human Endogenous Retroviruses (HERVs) that integrates both internal-region domain annotations and LTR regulatory architecture. The companion dataset containing conserved retroviral protein domains is released separately and is also part of HERVarium. 💻 Code and citation The full pipeline for LTR reconstruction, motif scanning, PBS/PPT detection, and structural annotation is available at:🔗 https://github.com/funcgen/herv-regulatory-map If you use this dataset in your research, please cite the dataset DOI and the corresponding manuscript (in preparation): Montserrat-Ayuso, T., & Esteve-Codina, A. (2025). Regulatory Features and Functional Specialization of Human Endogenous Retroviral LTRs: A Genome-Wide Annotation and Analysis via HERVarium. Manuscript in preparation. Version history v1: Initial release of genome-wide HERV LTR reconstructions, U3–R–U5 segments, motif annotations, and PBS/PPT predictions.