Description

This composite shows TAD domains from the 3D Genome Browser (3DGB) across 464 human Hi-C and Micro-C datasets on hg38. Each subtrack is one 3DGB dataset, displayed exactly as called and published by 3DGB. TAD domains are megabase-scale regions of the genome that preferentially self-interact; their boundaries (frequently bound by CTCF and cohesin) insulate neighboring regions and constrain enhancer-promoter contacts.

The 464 datasets span a wide range of normal and cancer samples, baseline and perturbation conditions, organs, and cell types, drawn from many published studies and re-processed by 3DGB through a single TAD-calling pipeline. They are browsable with a faceted selector (see below); the displayed domain intervals are 3DGB's own, with no UCSC re-calling, merging, lifting, or recurrence scoring.

Display Conventions and Configuration

Each subtrack is drawn as boxes spanning the self-interacting domains and is colored by organ. By default a small set of canonical reference datasets is shown (GM12878, H1-ESC, IMR-90, and HMEC); all other datasets are turned off and can be enabled through the faceted selector. Mousing over a domain shows the dataset name, organ, and assay.

These 464 datasets are not a cross-comparable consensus. Each represents one dataset's own TAD calls, made by different laboratories on different samples; coordinates are therefore not directly comparable across subtracks, and they are not directly comparable to the other TAD tracks in this set (which use different callers and resolutions). Because calls are made on binned Hi-C data (3DGB calls TADs at 25 kb), domain edges are uncertain to roughly the bin size, and domains do not tile the genome end to end.

Faceted selector

Use the faceted selector on the track configuration page to choose which datasets to display. Datasets can be filtered by:

• Organ – the organ of the sample (used for subtrack color).
• Cell type – the cell type, where annotated by 3DGB ("(unspecified)" when 3DGB does not record one).
• Assay – Hi-C or Micro-C.
• Condition – whether the sample is normal or cancer.
• Treatment – baseline (untreated) or perturbation (e.g. drug treatment, gene knockout, or other experimental manipulation).
• Provenance – see below.
• Year and Study – the publication year and the source data accession (e.g. a GEO series) that 3DGB re-processed.

The "Provenance" facet

Because this track ships every 3DGB human dataset, a few of the underlying source studies are already represented elsewhere in the UCSC Genome Browser. The Provenance facet flags these so they can be identified or filtered out:

• Novel to browser (422 datasets) – the source study is not otherwise displayed in the UCSC Genome Browser.
• Also in another UCSC track (42 datasets) – the underlying Hi-C study is already represented elsewhere in the UCSC Genome Browser, either as its own track or as an input to another track in this TAD set. These include Schmitt 2016 (11 datasets, also shown directly as the Schmitt 2016 boundaries track here), Rao 2014 (7 datasets; the Rao 2014 Hi-C maps are also offered in the browser), a few ENCODE datasets (overlapping the ENCODE contact domains track), and several studies (e.g. Dixon 2015) that are inputs to the TAD boundary stability track rather than displayed individually. The flag is intended to help avoid double-counting; it does not imply each dataset is separately viewable elsewhere.

To view only datasets that are new to the browser, select "Novel to browser" in the Provenance facet. Note that even the "Also in another UCSC track" datasets may differ in their displayed coordinates from the other UCSC tracks, because 3DGB re-processed and re-called each study through its own pipeline.

Methods

TAD domains were called by the 3D Genome Browser pipeline and are displayed verbatim. UCSC performed only a format normalization: each 3DGB per-dataset TAD file (a BED-like file with placeholder columns and an alternating two-color shading that carries no biological meaning) was reshaped to a plain four-column bigBed (chromosome, start, end, dataset name) and indexed. No domain coordinates were changed, and no re-calling, merging, lifting (all datasets are native hg38), or recurrence scoring was performed. The dataset metadata used to drive the faceted selector (organ, cell type, assay, year, study) was copied directly from the 3D Genome Browser.

Data Access

The raw data can be explored interactively with the Table Browser or the Data Integrator. For programmatic access, the track can be accessed using the Genome Browser's REST API. The underlying bigBed files can be downloaded from our download server. The complete original datasets are available from the 3D Genome Browser.

Credits

Thanks to the 3D Genome Browser team (Yue lab, Northwestern University) for assembling and uniformly processing these datasets. The 3D Genome Browser data are distributed under a CC BY-NC 4.0 license (free for non-commercial use). Please cite the 3D Genome Browser, and the original studies, when using these data.

References

Yu S, Fu Y, Wong JH, Wang J, Zhao H, Zhao J, Yue F. The 3D Genome Browser 2.0: an enhanced online platform for visualizing and analyzing 3D genome architecture. Nucleic Acids Res. 2026;54(D1):D48-D54. doi:10.1093/nar/gkaf1109

Wang Y, Song F, Zhang B, Zhang L, et al. The 3D Genome Browser: a web-based browser for visualizing 3D genome organization and long-range chromatin interactions. Genome Biol. 2018;19(1):151. doi:10.1186/s13059-018-1519-9