CoChat

Robust single-cell Hi-C clustering by convolution- and random-walk-based imputation

Open access
Zhou, Jingtian, Ma, Jianzhu, Chen, Yusi, Cheng, Chuankai, Bao, Bokan, Peng, Jian, Sejnowski, Terrence J., Dixon, Jesse R., Ecker, Joseph R. · 2019 · Proceedings of the National Academy of Sciences

The DOI matches its Crossref record; 42 of the 42 references it cites were found in a registry or an index.

Abstract

Three-dimensional genome structure plays a pivotal role in gene regulation and cellular function. Single-cell analysis of genome architecture has been achieved using imaging and chromatin conformation capture methods such as Hi-C. To study variation in chromosome structure between different cell types, computational approaches are needed that can utilize sparse and heterogeneous single-cell Hi-C data. However, few methods exist that are able to accurately and efficiently cluster such data into constituent cell types. Here, we describe scHiCluster, a single-cell clustering algorithm for Hi-C contact matrices that is based on imputations using linear convolution and random walk. Using both simulated and real single-cell Hi-C data as benchmarks, scHiCluster significantly improves clustering accuracy when applied to low coverage datasets compared with existing methods. After imputation by scHiCluster, topologically associating domain (TAD)-like structures (TLSs) can be identified within single cells, and their consensus boundaries were enriched at the TAD boundaries observed in bulk cell Hi-C samples. In summary, scHiCluster facilitates visualization and comparison of single-cell 3D genomes.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1073/pnas.1901423116

Cites

42 of 42 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1. The workflow of scHiCluster. The contact matrices of each single cell are smoothed by two steps of imputation that include convolution and random walk; these are based on the neighboring bins of a linear genome and long-range connections, respectively. To alleviate the coverage bias, only top 20% elements of the imputed matrices are selected. All single-cell matrices are then projected into the same space, and then clustering is performed to identify distinct cell types.
Fig. 1. The workflow of scHiCluster. The contact matrices of each single cell are smoothed by two steps of imputation that include convolution and random walk; these are based on the neighboring bins of a linear genome and long-range connections, respectively. To alleviate the coverage bias, only top 20% elements of the imputed matrices are selected. All single-cell matrices are then projected into the same space, and then clustering is performed to identify distinct cell types.CC BY-NC-ND · PubMed Central
Fig. 2. The embedding of simulated single cells at 1-Mb resolution from Rao et al. (29) (A) and Bonev et al. (30) (B) with different contact numbers. For each dataset, the embedding by scHiCluster is shown on the Left and by PCA is shown on the Right.
Fig. 2. The embedding of simulated single cells at 1-Mb resolution from Rao et al. (29) (A) and Bonev et al. (30) (B) with different contact numbers. For each dataset, the embedding by scHiCluster is shown on the Left and by PCA is shown on the Right.CC BY-NC-ND · PubMed Central
Fig. 3. The performance of scHiCluster and baseline methods on real single-cell Hi-C data. For Flyamer et al. (10) (A, C, and E) and Ramani et al. (7) (B, D, and F), the embedding (A and B) and ARI of clustering (C and D) are shown. For scHiCluster and eigenvector, the embeddings are shown in PC1 and PC2 space. For PCA and decay profile, the embedding is shown in PC2 and PC3 space. (E and F) The performance of scHiCluster and PCA on each chromosome (indicated by chr. numbers). The ARI using all chromosomes (indicated by star symbol).
Fig. 3. The performance of scHiCluster and baseline methods on real single-cell Hi-C data. For Flyamer et al. (10) (A, C, and E) and Ramani et al. (7) (B, D, and F), the embedding (A and B) and ARI of clustering (C and D) are shown. For scHiCluster and eigenvector, the embeddings are shown in PC1 and PC2 space. For PCA and decay profile, the embedding is shown in PC2 and PC3 space. (E and F) The performance of scHiCluster and PCA on each chromosome (indicated by chr. numbers). The ARI using all chromosomes (indicated by star symbol).CC BY-NC-ND · PubMed Central
Fig. 4. Visualization of contact matrices surrounding chr9:133,600,000–134,200,000 (the green box) in merged single-cell data and bulk data. The whole matrix shows a 2 M region from 132,900,000–134,900,000. The contact matrices were created by merging of 10 GM12878 cells (A) or K562 cells (B) after imputation. (C) The difference between A and B. SQRTVC normalized contact matrices from bulk GM12878 (D) and K562 (E) cell lines. (F) The difference between D and E. (G) Corresponding RNA-seq and H3K4me1 signals near genes from the indicated TAD region for both cell lines.
Fig. 4. Visualization of contact matrices surrounding chr9:133,600,000–134,200,000 (the green box) in merged single-cell data and bulk data. The whole matrix shows a 2 M region from 132,900,000–134,900,000. The contact matrices were created by merging of 10 GM12878 cells (A) or K562 cells (B) after imputation. (C) The difference between A and B. SQRTVC normalized contact matrices from bulk GM12878 (D) and K562 (E) cell lines. (F) The difference between D and E. (G) Corresponding RNA-seq and H3K4me1 signals near genes from the indicated TAD region for both cell lines.CC BY-NC-ND · PubMed Central
Fig. 5. scHiCluster facilitates identification of domain-like structures (TLSs). (A) The contact matrices at chr19:27.5M-29.5M of bulk Hi-C data with alternative TADs in ESC and NPC. (B) The downsampled contact matrices with 1-Mbp, 500-kbp, 250-kbp, and 100-kbp total contacts per cell before and after scHiCluster imputation. The green lines indicate the TLSs called from the plotted matrix, and the yellow lines represent the TADs called from bulk data of the corresponding cell type. (C) The number of downsampled ESCs with TLSs at chr19:28170000–28530000 and chr19:28530000–28770000, or downsampled NPCs with TLSs at chr19:28170000–28770000 being identified before and after scHiCluster imputation. (D) The number of single ESCs (1,007 in total) with TLSs boundary identified at each genome bin are shown by lines. The position of TADs boundaries identified in bulk data are presented by dots. The CTCF ChIP-seq signal from ENCODE is shown in the green track. The PC1 of bulk ESC Hi-C matrix is shown on the Top.
Fig. 5. scHiCluster facilitates identification of domain-like structures (TLSs). (A) The contact matrices at chr19:27.5M-29.5M of bulk Hi-C data with alternative TADs in ESC and NPC. (B) The downsampled contact matrices with 1-Mbp, 500-kbp, 250-kbp, and 100-kbp total contacts per cell before and after scHiCluster imputation. The green lines indicate the TLSs called from the plotted matrix, and the yellow lines represent the TADs called from bulk data of the corresponding cell type. (C) The number of downsampled ESCs with TLSs at chr19:28170000–28530000 and chr19:28530000–28770000, or downsampled NPCs with TLSs at chr19:28170000–28770000 being identified before and after scHiCluster imputation. (D) The number of single ESCs (1,007 in total) with TLSs boundary identified at each genome bin are shown by lines. The position of TADs boundaries identified in bulk data are presented by dots. The CTCF ChIP-seq signal from ENCODE is shown in the green track. The PC1 of bulk ESC Hi-C matrix is shown on the Top.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 42 references

Cited by

Items 101–183 of 183

Metadata

Authors
Zhou, J., Ma, J., Chen, Y., Cheng, C., Bao, B., Peng, J., Sejnowski, T. J., Dixon, J. R., & Ecker, J. R.
Affiliations
  • Salk Institute for Biological Studies (US)
  • University of California San Diego (US)
  • University of Illinois Urbana-Champaign (US)
  • Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037;
Corresponding author
Ecker, Joseph R.
Venue
Proceedings of the National Academy of Sciences
Published
2019-06-24
Volume
116
Issue
28
Pages
14011-14018
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
OpenAlex
W2949773289
Cited by
180
References
42
Citations by year
2026: 29 · 2025: 34 · 2024: 27 · 2023: 20 · 2022: 27 · 2021: 31 · 2020: 9 · 2019: 3
FWCI
6.12
Type
journal article
Language
en
Open access
Locations
  • Proceedings of the National Academy of Sciences (publishedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
Funders
  • Howard Hughes Medical Institute
  • National Human Genome Research Institute (R21HG009274)
  • National Cancer Institute (P30CA014195)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genomic variations and chromosomal abnormalities
Keywords
single-cell Hi-C, 3D genome organization, chromatin conformation, cell type clustering, linear convolution, random walks, topologically associating domains, TAD boundaries, data sparsity, low coverage, chromatin structure
MeSH terms
Algorithms, Chromatin, Humans, Molecular Conformation, Cluster Analysis, Genome, Computational Biology, Chromosome Structures, Single-Cell Analysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.