CoChat

ChIP-Enrich: gene set enrichment testing for ChIP-seq data

Open access

The DOI matches its Crossref record; 52 of the 52 references it cites were found in a registry or an index.

Abstract

Gene set enrichment testing can enhance the biological interpretation of ChIP-seq data. Here, we develop a method, ChIP-Enrich, for this analysis which empirically adjusts for gene locus length (the length of the gene body and its surrounding non-coding sequence). Adjustment for gene locus length is necessary because it is often positively associated with the presence of one or more peaks and because many biologically defined gene sets have an excess of genes with longer or shorter gene locus lengths. Unlike alternative methods, ChIP-Enrich can account for the wide range of gene locus length-to-peak presence relationships (observed in ENCODE ChIP-seq data sets). We show that ChIP-Enrich has a well-calibrated type I error rate using permuted ENCODE ChIP-seq data sets; in contrast, two commonly used gene set enrichment methods, Fisher's exact test and the binomial test implemented in Genomic Regions Enrichment of Annotations Tool (GREAT), can have highly inflated type I error rates and biases in ranking. We identify DNA-binding proteins, including CTCF, JunD and glucocorticoid receptor α (GRα), that show different enrichment patterns for peaks closer to versus further from transcription start sites. We also identify known and potential new biological functions of GRα. ChIP-Enrich is available as a web interface (http://chip-enrich.med.umich.edu) and Bioconductor package.

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 · 10.1093/nar/gku463

Cites

52 of 52 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. Gene locus length-to-peak presence relationship becomes stronger as the total number of peaks increases. The relationship between gene locus length and proportion of genes with ≥1 peak in a gene locus varies widely in 63 ENCODE ChIP-seq data sets, from no relationship to strongly positive. DNA-binding proteins (DBPs) from the GM12878 cell line were categorized into three groups of 21 DBPs by the total number of peaks. For each DBP, the relationship between log10 locus length and proportion of genes with a peak was modeled using a binomial cubic smoothing spline (see the Materials and Methods section). (a)–(c) Bar plots show the average proportion of peaks present within the specified distance from the TSS (kb) (gray bar) and the proportions for individual DBPs (colored dots, the same color as the line in the corresponding plot). DBPs with fewer peaks tend to have a higher proportion of binding close to TSSs. (d) The locus length-to peak presence relationship tends to be weak for data sets with few peaks. (e) and (f) The relationship becomes strongest when the number of peaks is highest (f). None of the DBPs in (d), two of the DBPs in (e) and 10 of the DBPs in (f) are histone modifications.
Figure 1. Gene locus length-to-peak presence relationship becomes stronger as the total number of peaks increases. The relationship between gene locus length and proportion of genes with ≥1 peak in a gene locus varies widely in 63 ENCODE ChIP-seq data sets, from no relationship to strongly positive. DNA-binding proteins (DBPs) from the GM12878 cell line were categorized into three groups of 21 DBPs by the total number of peaks. For each DBP, the relationship between log10 locus length and proportion of genes with a peak was modeled using a binomial cubic smoothing spline (see the Materials and Methods section). (a)–(c) Bar plots show the average proportion of peaks present within the specified distance from the TSS (kb) (gray bar) and the proportions for individual DBPs (colored dots, the same color as the line in the corresponding plot). DBPs with fewer peaks tend to have a higher proportion of binding close to TSSs. (d) The locus length-to peak presence relationship tends to be weak for data sets with few peaks. (e) and (f) The relationship becomes strongest when the number of peaks is highest (f). None of the DBPs in (d), two of the DBPs in (e) and 10 of the DBPs in (f) are histone modifications.CC BY · PubMed Central
Figure 2. Overview of ChIP-Enrich. We describe ChIP-Enrich in four steps. (1) ChIP-seq peaks are assigned to genes using a chosen gene locus definition. Definitions include: ‘nearest gene’, ‘≤1kb from TSS’ and ‘nearest TSS’. (2) It is determined whether ≥1 peak is present in each gene locus. (3) Gene set enrichment is performed for each gene set using a logistic regression model, adjusting for locus length with a binomial cubic smoothing spline term (represented as f in the model equation). (4) Data and results are summarized. (a) Plot of observed spline fit for log10 locus length versus proportion of genes with a peak (orange). Expected line if no relationship between log10 locus length and proportion of genes with a peak (dark gray, satisfies Fisher's exact test assumptions). Expected line if the number of peaks observed is proportional to locus length (light gray, binomial test assumption). For visualization only, each point is the proportion of genes that are assigned a peak within sequential bins of 25 genes. (b) Bar plot of the proportion of peaks found at various distances from the TSS. (c) Abbreviated ChIP-Enrich output.
Figure 2. Overview of ChIP-Enrich. We describe ChIP-Enrich in four steps. (1) ChIP-seq peaks are assigned to genes using a chosen gene locus definition. Definitions include: ‘nearest gene’, ‘≤1kb from TSS’ and ‘nearest TSS’. (2) It is determined whether ≥1 peak is present in each gene locus. (3) Gene set enrichment is performed for each gene set using a logistic regression model, adjusting for locus length with a binomial cubic smoothing spline term (represented as f in the model equation). (4) Data and results are summarized. (a) Plot of observed spline fit for log10 locus length versus proportion of genes with a peak (orange). Expected line if no relationship between log10 locus length and proportion of genes with a peak (dark gray, satisfies Fisher's exact test assumptions). Expected line if the number of peaks observed is proportional to locus length (light gray, binomial test assumption). For visualization only, each point is the proportion of genes that are assigned a peak within sequential bins of 25 genes. (b) Bar plot of the proportion of peaks found at various distances from the TSS. (c) Abbreviated ChIP-Enrich output.CC BY · PubMed Central
Figure 3. Representative plots of the four patterns of enrichment comparing the ‘≤1kb from TSS and nearest TSS’ locus definitions. Gene set enrichment testing using the ‘≤1kb from TSS’ and ‘nearest TSS’ locus definitions may identify similar (a) or different (b) sets of significant GO terms for the same DBP. Alternatively, most of the enrichment signal may come from ‘nearest TSS’ that uses all peaks (c) or ‘≤1 kb from TSS’ that ignores peaks >1 kb from a TSS (d). (a)–(d) Upper plot: bar plot of the proportion of peaks at different distances from the TSS. Lower plot: comparison of –log10(P-values) from ChIP-Enrich GO term enrichment testing using ‘≤1 kb from TSS’ versus ‘nearest TSS’ locus definitions in ENCODE data for the GM12878 cell line. GO terms enriched with FDR ≤ 0.05 for: ‘≤1 kb from TSS’ only (green); ‘nearest TSS’ only (blue); ‘≤1 kb from TSS’ and ‘nearest TSS’ (orange); neither analysis (black). r, Pearson correlation coefficient. These patterns are representative of patterns present in 63 ENCODE DBPs from the GM12878 cell line.
Figure 3. Representative plots of the four patterns of enrichment comparing the ‘≤1kb from TSS and nearest TSS’ locus definitions. Gene set enrichment testing using the ‘≤1kb from TSS’ and ‘nearest TSS’ locus definitions may identify similar (a) or different (b) sets of significant GO terms for the same DBP. Alternatively, most of the enrichment signal may come from ‘nearest TSS’ that uses all peaks (c) or ‘≤1 kb from TSS’ that ignores peaks >1 kb from a TSS (d). (a)–(d) Upper plot: bar plot of the proportion of peaks at different distances from the TSS. Lower plot: comparison of –log10(P-values) from ChIP-Enrich GO term enrichment testing using ‘≤1 kb from TSS’ versus ‘nearest TSS’ locus definitions in ENCODE data for the GM12878 cell line. GO terms enriched with FDR ≤ 0.05 for: ‘≤1 kb from TSS’ only (green); ‘nearest TSS’ only (blue); ‘≤1 kb from TSS’ and ‘nearest TSS’ (orange); neither analysis (black). r, Pearson correlation coefficient. These patterns are representative of patterns present in 63 ENCODE DBPs from the GM12878 cell line.CC BY · PubMed Central
Figure 4. Comparison of GRα enrichment results for ChIP-seq (using two locus definitions) and RNA-seq data from A549 cells. Enriched GO terms for differentially expressed transcripts and GRα binding following 100-nM DEX treatment show stronger overlap using the ‘nearest TSS’ locus definition than using the ‘≤1 kb from TSS’ definition. (a) Observed spline fit for GRα fits neither FET nor the binomial test assumption (orange); bar plot of the proportion of peaks at different distances from the TSS. See Figure 2(4a) and (b) for further details. (b) Using the ‘nearest TSS’ locus definition with GRα results in more overlapping terms with RNA-seq results than using ‘≤1 kb from TSS’. (c) Using the top 195 ranked terms for each test, FET and the binomial test have more overlap with ChIP-Enrich than with each other. (d)–(f) Comparison of –log10(P-values) for GO term enrichment tests based on ChIP-seq data (ChIP-Enrich) and/or RNA-seq (GOseq) data. (f) Many enriched RNA-seq terms would have been missed in the ChIP-seq data if only peaks in promoter regions were considered. GO terms enriched and FDR ≤ 0.05: for Y-axis test only (green); for X-axis test only (blue); for X- and Y-axis tests (orange); for neither (black). Vasculature development and related GO terms (triangles). The majority of GO terms that overlap between ‘≤1 kb from TSS’ and ‘nearest TSS’ are related to fatty acid metabolism, reactive oxygen species and unfolded proteins, or blood coagulation.
Figure 4. Comparison of GRα enrichment results for ChIP-seq (using two locus definitions) and RNA-seq data from A549 cells. Enriched GO terms for differentially expressed transcripts and GRα binding following 100-nM DEX treatment show stronger overlap using the ‘nearest TSS’ locus definition than using the ‘≤1 kb from TSS’ definition. (a) Observed spline fit for GRα fits neither FET nor the binomial test assumption (orange); bar plot of the proportion of peaks at different distances from the TSS. See Figure 2(4a) and (b) for further details. (b) Using the ‘nearest TSS’ locus definition with GRα results in more overlapping terms with RNA-seq results than using ‘≤1 kb from TSS’. (c) Using the top 195 ranked terms for each test, FET and the binomial test have more overlap with ChIP-Enrich than with each other. (d)–(f) Comparison of –log10(P-values) for GO term enrichment tests based on ChIP-seq data (ChIP-Enrich) and/or RNA-seq (GOseq) data. (f) Many enriched RNA-seq terms would have been missed in the ChIP-seq data if only peaks in promoter regions were considered. GO terms enriched and FDR ≤ 0.05: for Y-axis test only (green); for X-axis test only (blue); for X- and Y-axis tests (orange); for neither (black). Vasculature development and related GO terms (triangles). The majority of GO terms that overlap between ‘≤1 kb from TSS’ and ‘nearest TSS’ are related to fatty acid metabolism, reactive oxygen species and unfolded proteins, or blood coagulation.CC BY · 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 52 references

Cited by

See all 164

Metadata

Authors
Welch, R. P., Lee, C., Imbriano, P. M., Patil, S., Weymouth, T. E., Smith, R. A., Scott, L. J., & Sartor, M. A.
Affiliations
University of Michigan (US)
Corresponding author
  • Scott, Laura J.
  • Sartor, Maureen A.
Venue
Nucleic Acids Research
Published
2014-05-30
Volume
42
Issue
13
Pages
e105-e105
ISSN
  • 0305-1048 (ISSN-L)
  • 1362-4962
OpenAlex
W2168131180
Cited by
151
References
52
Citations by year
2026: 10 · 2025: 11 · 2024: 19 · 2023: 16 · 2022: 14 · 2021: 10 · 2020: 16 · 2019: 16 · 2018: 14 · 2017: 4 · 2016: 11 · 2015: 9 · 2014: 1
FWCI
3.42
Type
journal article
Language
en
Open access
Locations
  • Nucleic Acids Research (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
Funders
  • National Institutes of Health (ES017885, DK062370)
  • National Human Genome Research Institute (R56HG000376, R01HG000376, T32HG000040)
  • National Cancer Institute (T32CA140044, R01CA158286)
  • National Institute of Diabetes and Digestive and Kidney Diseases (DK-062370, R01DK062370)
  • National Institute of Environmental Health Sciences (P30ES017885)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
Keywords
gene set enrichment analysis, ChIP-seq, type I error rate, GREAT, Fisher's exact test, binomial test, DNA-binding proteins, CTCF, JunD, glucocorticoid receptor alpha, transcription start sites, peak calling, ENCODE
MeSH terms
DNA-Binding Proteins, Genes, Receptors, Glucocorticoid, Logistic Models, Sequence Analysis, DNA, Chromatin Immunoprecipitation, Genetic Loci

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: ChIP-Enrich: gene set enrichment testing for ChIP-seq data. https://app.cochat.ai/papers/10.1093/nar/gku463
The paper
Welch, R. P., Lee, C., Imbriano, P. M., Patil, S., Weymouth, T. E., Smith, R. A., Scott, L. J., & Sartor, M. A. (2014). ChIP-Enrich: gene set enrichment testing for ChIP-seq data. Nucleic Acids Research. https://doi.org/10.1093/nar/gku463

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.