Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Feb 5;116(6):2175-2180.
doi: 10.1073/pnas.1808631116. Epub 2019 Jan 18.

Chromatin features constrain structural variation across evolutionary timescales

Geoff Fudenberg  1 , Katherine S Pollard  1   2   3   4   5   6
Affiliations

Chromatin features constrain structural variation across evolutionary timescales

Geoff Fudenberg et al. Proc Natl Acad Sci U S A. .

Abstract

The potential impact of structural variants includes not only the duplication or deletion of coding sequences, but also the perturbation of noncoding DNA regulatory elements and structural chromatin features, including topological domains (TADs). Structural variants disrupting TAD boundaries have been implicated both in cancer and developmental disease; this likely occurs via "enhancer hijacking," whereby removal of the TAD boundary exposes enhancers to new target transcription start sites (TSSs). With this functional role, we hypothesized that boundaries would display evidence for negative selection. Here we demonstrate that the chromatin landscape constrains structural variation both within healthy humans and across primate evolution. In contrast, in patients with developmental delay, variants occur remarkably uniformly across genomic features, suggesting a potentially broad role for enhancer hijacking in human disease.

Keywords: CTCF; Hi-C; chromatin; comparative genomics; evolution.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
Approach to detect purifying selection against deleterious structural variants. (A) To study sets of structural variants subject to purifying selection for varying amounts of time, we obtained structural variants representing divergence with great apes (30), variation within the human population (31), and detected in patients (shown with red crosses) with developmental delay and autism (31). (B) To characterize the chromatin landscape, we curated: chromatin states, TSSs, CTCF binding clusters, and TAD boundaries. (C) We summarize each set of variants by their breakpoint frequency and coverage across the genome. (D) We then determine whether genomic features are relatively enriched or depleted for variant breakpoints and coverage. As structural variants subject to purifying selection are gradually removed from the population over time, we expect features under purifying selection to be depleted for breakpoint frequency and coverage.
Fig. 2.
Fig. 2.
Ape deletions show patterns of purifying selection at active chromatin states, CTCF clusters, and TAD boundaries. (A) Deletions observed in apes (30) have relatively low coverage and breakpoint frequency in active genomic features and at TAD boundaries. Circles represent the average across 127 Roadmap cell types; see SI Appendix, Fig. S1A for variability of these estimates across cell types. Log10(observed/expected) represents deviations from a uniform distribution across the genome, accounting for the proportion of the genome covered by a given genomic feature (Methods). State 3 had no observed breakpoints or coverage and is shown with a black center at the minimal plotted x–y value, for display. (B) Ape deletion coverage at TSSs (Left) and CTCF clusters (Right) scales with the strength of these genomic features. Curves show average expected coverage as a function of feature strength in a sliding window (±5 percentiles); shaded areas represent 5th and 95th percentiles calculated over 1,000 bootstrap samples.
Fig. 3.
Fig. 3.
Human deletions reveal the spectrum of purifying selection across genomic features. (A) Deletions observed in healthy humans (31) have lower coverage and breakpoint frequency in active states and at TAD boundaries. Circles represent the average across 127 Roadmap cell types. Log10(observed/expected) represents deviations from a uniform distribution across the genome, as in Fig. 2 (Methods). (B) Healthy human deletion coverage at TSSs (Left) and CTCF clusters (Right) scales with the strength of these genomic features, plotted as in Fig. 2B. (C–F) Coverage in ±500-kb genomic region at 10-kb binned resolution. (C) TAD boundaries shared across cell types are more depleted for human deletions than those found in only one cell type. (D) TAD boundaries with more evolutionary conservation at the base-pair level are more depleted for deletions. (E) Deletions shared across individuals are more depleted at TAD boundaries. (F) TAD boundaries are more depleted for deletions than Hi-C peak bases. Note these and other curves approach zero at ∼5–10 Mb (SI Appendix, Fig. S3H).
Fig. 4.
Fig. 4.
Deletions in human disease show little avoidance of key genomic features. (A) Average deletion coverage and breakpoint frequency for TSSs stratified and shaded by strength. Log10(observed/expected) represents deviations from a uniform distribution across the genome, as in Figs. 2 and 3 (Methods). Unlike for healthy subjects (blue), deletions from patients with developmental delay and autism (orange) show no avoidance of strong TSSs, either for coverage or breakpoint frequency, both cohorts from Coe et al. (31). (B) Deletion coverage in patients shows little relationship with TSS strength. (C and D) As for A and B, but for CTCF clusters. (E) Binned deletion coverage at 100 kb from patients (orange) and healthy controls (blue) across the first four chromosomes illustrate differences in their large-scale distribution across the genome. (F) Binned deletion coverage at 10 kb above tracks showing inactive (gray) versus active (green) Roadmap states. The region on chr1 with a long stretch of inactive states (Left) shows an island of high coverage in healthy subjects; the mixed states on chr3 (Right) show broadly elevated coverage in patients, compared with the more punctuated coverage in healthy subjects. (G) Binned deletion coverage at 100 kb, colored by GC content and shown on a log scale with a pseudocount of one for display, do not highly correspond for patients and healthy subjects. (H) Coverage per 100-kb bin shows more uniformity in patients. (I) Autocorrelation of 100-kb binned deletion coverage profiles varies more slowly in patients.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Zhang F, Lupski JR. Non-coding genetic variants in human disease. Hum Mol Genet. 2015;24:R102–R110. - PMC - PubMed
    1. Chiang C, et al. GTEx Consortium The impact of structural variation on human gene expression. Nat Genet. 2017;49:692–699. - PMC - PubMed
    1. Cheng Z, et al. A genome-wide comparison of recent chimpanzee and human segmental duplications. Nature. 2005;437:88–93. - PubMed
    1. Krijger PHL, de Laat W. Regulation of disease-associated gene expression in the 3D genome. Nat Rev Mol Cell Biol. 2016;17:771–782. - PubMed
    1. Spielmann M, Mundlos S. Looking beyond the genes: The role of non-coding variants in human disease. Hum Mol Genet. 2016;25:R157–R165. - PubMed

Publication types

LinkOut - more resources