Description
Fiber-seq maps which parts of each DNA molecule are bound by proteins and which are open.
Permeabilized cells are treated with a non-specific adenine methyltransferase that methylates
only exposed adenines, so nucleosomes and other bound proteins leave an unmethylated footprint
on each molecule. The DNA is then sequenced on a PacBio instrument, which reads the base
sequence, the added adenine methylation and the native CpG methylation from the same molecule.
Because the reads are long enough to span heterozygous variants, they can be phased to a
parental haplotype, giving a chromatin picture for each allele separately rather than an average
of the two.
The accessibility this measures is the same property that DNase-seq and ATAC-seq measure, and
the tracks here can be read much as those are. The difference is that Fiber-seq reports it one
molecule at a time, so a value is a fraction of molecules rather than a pile of cut sites, and
it can be split by haplotype.
This collection holds the Fiber-seq data from the Stergachis and Vollger labs for 41 samples,
covering common cell lines, 20 lymphoblastoid lines from individuals sequenced by the Human
Pangenome Reference Consortium, and five rare disease cases. It contains two tracks:
- Fiber-seq Accessible shows percent-accessible chromatin
for seven widely used cell lines as a single overlay, so accessibility can be compared
across cell types at a glance.
- Fiber-seq Compendium holds everything, for
every one of the 41 samples: the percent-accessible signal, the called FIRE regulatory
element peaks, CpG methylation, and per-haplotype versions of the accessibility and the
methylation. Samples are selected from a searchable table.
Related data can be found in the DNA Methylation
collection, which holds methylation measured by other assays.
Display Conventions
Both tracks have their own display conventions and color schemes; see the individual track
description pages linked above. The accessibility and methylation signals are scaled 0 to 100
percent throughout, so a bar height means the same thing in every track and every sample. The
exception is the CpG haplotype difference, which is a difference between two percentages and so
runs from -100 to +100.
Methods
The data were generated and processed by the Stergachis and Vollger labs. Details of the assay,
the machine learning model that calls regulatory elements from single molecules, and the
methylation calling are given on the individual track description pages.
Data Access
The individual track description pages linked above explain how to download the underlying
bigWig and bigBed files and how to query them from scripts.
Credits
Thanks to Mitchell Vollger, Andrew Stergachis and Shane Neph at the University of Washington
and the University of Utah for generating this data, for packaging it for the browser and for
their help in arranging the tracks.
References
Vollger MR, Swanson EG, Neph SJ, Ranchalis J, Munson KM, Ho CH, Cheng YHH, Sedeño-Cortés AE, Fondrie
WE, Bohaczuk SC et al.
A haplotype-resolved view of human gene regulation.
bioRxiv. 2025 Jun 2;.
PMID: 40501892; PMC: PMC12157683
Stergachis AB, Debo BM, Haugen E, Churchman LS, Stamatoyannopoulos JA.
Single-molecule regulatory architectures captured by chromatin fiber sequencing.
Science. 2020 Jun 26;368(6498):1449-1454.
PMID: 32587015