Description

This track shows structural variants (SVs) from the third phase of the Human Genome Structural Variation Consortium (HGSVC3). The callset comes from 65 diverse individuals across five continental groups, each sequenced with PacBio HiFi (~47x), Oxford Nanopore ultra-long reads (~56x) and complemented with Strand-seq, optical mapping, Hi-C and Iso-Seq for haplotype-resolved assembly. SVs were discovered from the de novo assemblies with PAV v2.4.0.1 and cross-validated by ten additional orthogonal callers.

The track merges the two final SV annotation tables from the HGSVC3 v1.0 release on GRCh38: 176,232 insertions/deletions and 300 inversions, for a total of 176,532 SVs. Each row is a site-level variant with the list of carrier haplotypes and additional structural annotations.

The same track is also available natively on the T2T-CHM13 (hs1) assembly: HGSVC3 independently aligned all haplotype-resolved assemblies to both GRCh38 and T2T-CHM13 and released a separate set of annotation tables per reference. The hs1 track is built directly from the HGSVC3 T2T-CHM13 annotation tables (188,224 DEL+INS and 276 INV; 188,500 SVs total) — no liftOver is involved.

Display Conventions and Configuration

Items are colored by SV type:

• Deletions (DEL) - red
• Insertions (INS) - blue
• Inversions (INV) - orange

Insertions are placed at the insertion site with a width of 1 bp; deletions and inversions span the affected reference interval. Filters are available for SV type, SV length, carrier-haplotype count, distinct sample count, whether the site falls in a Tandem Repeat Finder region and the fraction of the variant overlapping segmental duplications.

The detail page shows, where available:

• Allele / Sample Count: number of carrier haplotypes (out of the 2*65 = 130 phased haplotypes plus unphased "un" entries) and the number of distinct samples carrying the variant.
• Reference / Contig Homology: microhomology length (5',3') at the breakpoints in the reference and in the assembly contig (insertions and deletions only).
• Inner Inversion Region: for inversions, the coordinate range of the inner inverted sequence, distinct from the outer breakpoint interval.
• Transposable Element: when the inserted or deleted sequence was classified as a known TE family.
• Segmental Duplication Overlap: fraction of the variant interval overlapping UCSC segmental duplications in the reference.
• Carrier Haplotypes: full list of haplotype IDs (e.g. HG00096-h1, HG00096-h2, HG00514-un) carrying the variant.

Methods

Logsdon et al. 2025 produced fully phased hybrid de novo assemblies for 65 diverse individuals (63 from 1kGP, NA21487 from HapMap, and HG002 from GIAB), using PacBio HiFi (Sequel II/Revio, 30-h movies), Oxford Nanopore ultra-long sequencing (R9.4.1 PromethION, 96-h runs), Bionano optical mapping (DLE-1 on Saphyr 2nd-gen), Strand-seq, Hi-C (Proximo) and Iso-Seq. Assemblies were generated with Verkko v1.4.1 (primary) and hifiasm-UL v0.19.6 (complementary, especially for centromeres and Yq12), phased with the Graphasing pipeline v0.3.1-alpha, and produced 130 haplotype assemblies with median N50 of 130 Mbp that close 92% of previous assembly gaps (39% of chromosomes at telomere-to-telomere status). SVs were called against GRCh38 and T2T-CHM13 with PAV v2.4.1 (plus DipCall and SVIM-asm from the same alignments) and cross-validated with an additional ten callers (PBSV, Sniffles, Delly, cuteSV, DeBreak, SVIM, DeepVariant, Clair3, PEPPER-Margin-DeepVariant for ONT and MELT-LRA/PALMER2 for MEIs). Calls were merged with SV-Pop and centromere-satellite / telomere hits were filtered. The final GRCh38 release contains 176,232 DEL+INS plus 300 INV (176,532 SVs total); the T2T-CHM13 release contains 188,224 DEL+INS plus 276 INV (188,500 SVs total).

For display, the two final HGSVC3 v1.0 annotation tables variants_GRCh38_sv_insdel_HGSVC2024v1.0.tsv.gz and variants_GRCh38_sv_inv_HGSVC2024v1.0.tsv.gz were downloaded from the IGSR HGSVC3 GRCh38 release directory and merged into a single bigBed. The hs1 version uses the parallel variants_T2T-CHM13_sv_insdel_HGSVC2024v1.0.tsv.gz and variants_T2T-CHM13_sv_inv_HGSVC2024v1.0.tsv.gz tables from the HGSVC3 T2T-CHM13 release directory; no liftOver is involved on hs1. Type-specific columns (HOM_REF/HOM_TIG/TE for insdel; RGN_REF_INNER for inversions) are empty on the detail page when they do not apply.

The step-by-step build commands (download, format conversion, bigBed build) are recorded in the UCSC makeDoc for this track container: doc/hg38/lrSv.txt and doc/hs1/lrSv.txt. The conversion scripts and autoSql schemas live in makeDb/scripts/lrSv.

Data Access

The data can be explored interactively in table format with the Table Browser or the Data Integrator, and accessed programmatically through our API, track=hgsvc3Sv.

The bigBed is available from our download server for both assemblies:

• GRCh38: hg38 hgsvc3.bb
• T2T-CHM13: hs1 hgsvc3.bb

The original annotation tables are available from the HGSVC3 release on the IGSR FTP site.

Credits

Thanks to the Human Genome Structural Variation Consortium (HGSVC) and all participating sequencing and analysis centers for making the HGSVC3 annotation tables publicly available.

References

Logsdon GA, Ebert P, Audano PA, Loftus M, Porubsky D, Ebler J, Yilmaz F, Hallast P, Prodanov T, Yoo D et al. Complex genetic variation in nearly complete human genomes. Nature. 2025 Aug;644(8076):430-441. PMID: 40702183; PMC: PMC12350169