A typical enhancer discovery project ends with a list: dozens to hundreds of candidate enhancers — from ATAC-seq peaks, ChIP-seq H3K27ac regions, or GWAS fine-mapped variants — each suspected of regulating something, somewhere, at some distance. The next question is the hard one: which gene does each candidate regulate?

Linear proximity is not a reliable proxy. Approximately 40–60% of confirmed enhancer-promoter interactions skip the nearest gene, and regulatory contacts can span hundreds of kilobases, bypassing multiple intervening promoters. Assigning a candidate enhancer to its target gene therefore requires direct measurement of three-dimensional chromatin contacts. The question is which method to use.

This article compares the three methods most commonly applied to enhancer-target gene mapping — HiChIP, Capture Hi-C, and Micro-C — and provides a decision framework for choosing among them based on your experimental context: what you already know, what you need to prove, and what resources you have available.

Figure 1: Three methods for enhancer-promoter interaction mapping — HiChIP enriches by protein/mark, Capture Hi-C enriches by genomic region, and Micro-C provides unbiased nucleosome-resolution maps.

Why Assigning Enhancers to Target Genes Requires Chromatin Interaction Data

Enhancers regulate promoters through physical proximity in three-dimensional nuclear space, not through linear genomic distance. A regulatory element 200 kb away from a gene can be its primary activator while the nearest promoter 20 kb away is entirely unaffected. This is not an edge case — it is a common mode of enhancer function that linear annotation systematically misassigns.

Three experimental strategies have been used historically to link enhancers to genes:

Proximity-based assignment assigns each enhancer to the nearest transcription start site. It is simple but wrong often enough to misdirect validation resources. In the CRISPRi survey by Fulco et al. (2019), linear distance alone was a poor predictor of functional enhancer-gene relationships, and the Activity-by-Contact (ABC) model — which incorporates Hi-C contact frequency — substantially outperformed distance-only methods.

Correlation-based assignment links enhancers to genes by correlating chromatin accessibility or histone modification signal across samples with target gene expression. This approach requires matched epigenomic and transcriptomic data from multiple conditions or cell types and performs poorly when enhancer activity and gene expression are not linearly coupled, which is common in steady-state conditions.

Contact-based assignment directly measures which DNA sequences physically interact in three-dimensional space using chromatin conformation capture. This is the gold standard for enhancer-target gene assignment because it measures the physical proximity that underlies enhancer function, rather than inferring it from genomic distance or expression correlation.

Three Methods for Enhancer-Promoter Interaction Mapping

Each method captures chromatin contacts differently — by what it enriches, what it ignores, and at what resolution it resolves interacting regions. The choice among them depends on whether your starting point is a set of candidate regions, a protein or histone mark of interest, or an unbiased genome-wide survey. For a broader comparison that also includes standard Hi-C, see the guide to Hi-C, Micro-C, and Capture Hi-C method selection.

HiChIP: Protein-Centric Interaction Mapping

HiChIP, developed by Mumbach et al. (2016), combines chromatin immunoprecipitation with in situ Hi-C proximity ligation. An antibody against a chromatin-associated protein — typically a histone modification such as H3K27ac marking active enhancers and promoters, or a structural protein such as cohesin or CTCF — enriches chromatin fragments carrying that mark before proximity ligation. The resulting library captures contacts where at least one anchor carries the targeted modification.

The method's defining characteristic is efficiency. Because the immunoprecipitation step concentrates sequencing on interaction fragments involving the protein of interest, HiChIP requires approximately 10-fold fewer reads than equivalent-resolution Hi-C to resolve regulatory contacts. H3K27ac HiChIP specifically enriches for interactions involving active regulatory elements — precisely the contacts most relevant to enhancer-target gene assignment — while filtering out the structural and inactive chromatin contacts that dominate unenriched Hi-C libraries.

The trade-off is scope. HiChIP reports contacts for regions carrying the targeted mark. Contacts involving enhancers or promoters not marked by H3K27ac — including poised, repressive, or tissue-specific regulatory elements marked by other modifications — are invisible. If your biological question concerns a specific transcription factor or histone mark, this selectivity is an advantage. If you need an unbiased map of all possible regulatory contacts at a locus, it is a limitation.

Capture Hi-C: Region-Centric Interaction Mapping

Capture Hi-C, first applied to promoter interactions by Mifsud et al. (2015), starts with a standard Hi-C library and enriches for contacts involving predefined genomic regions using hybridization-based capture with biotinylated RNA or DNA probes. The capture regions can be promoters (promoter Capture Hi-C), GWAS-identified variants, candidate enhancers, or any custom genomic interval.

The defining advantage of Capture Hi-C is density. By targeting sequencing to a few thousand to tens of thousands of loci, Capture Hi-C achieves high-resolution contact maps at regions of interest at a fraction of the sequencing cost of genome-wide methods. In the original promoter Capture Hi-C study, Mifsud et al. profiled contacts for approximately 22,000 promoters in two human blood cell types, identifying over 1.6 million promoter interactions. These data revealed that interacting regions are enriched for disease-associated SNPs from GWAS, providing a mechanistic framework for non-coding variant interpretation.

Capture Hi-C is best suited when you have a predefined list of regions — enhancer candidates from ATAC-seq or ChIP-seq, fine-mapped GWAS loci, or all annotated promoters — and you want saturated interaction maps at those specific loci. The method requires a bespoke probe set designed for your regions of interest, which introduces an upfront design cost and time investment. Probe design quality — uniform coverage across target regions, balanced GC content, and adequate tiling density — directly determines capture efficiency and data uniformity.

Micro-C: Unbiased, Nucleosome-Resolution Interaction Mapping

Micro-C, originally developed in yeast by Hsieh et al. (2015) and extended to mammalian cells by Hsieh et al. (2020) and Krietenstein et al. (2020), replaces restriction enzyme fragmentation with micrococcal nuclease (MNase) digestion to generate chromatin fragments at mononucleosome resolution. The resulting contact maps resolve interactions at the scale of individual nucleosomes — approximately 150–200 bp — rather than the 1–10 kb bins typical of standard Hi-C.

The method's key output for enhancer biology is the detection of fine-scale loops that Hi-C and even HiChIP miss. Krietenstein et al. identified approximately 20,000 additional loops in human cells that were invisible to Hi-C, many of which corresponded to enhancer-promoter contacts at actively transcribed genes. Micro-C also resolves the internal structure of interaction domains — including the stripe-shaped contact patterns at super-enhancers and the precise positioning of domain boundaries relative to CTCF sites and transcription start sites.

Micro-C is the method of choice when you need an unbiased, genome-wide map of chromatin contacts at the highest achievable resolution. The trade-offs are sequencing depth — Micro-C requires deeper sequencing than Hi-C to achieve nucleosome resolution genome-wide — and the absence of protein- or region-specific enrichment, meaning that most sequencing reads report structural rather than regulatory contacts. If your goal is to discover which regulatory elements contact which promoters without prior assumptions about which marks or regions to target, Micro-C provides the most complete picture.

Method Selection Framework

The right method depends on what you already have — a list of regions, a protein of interest, or neither — and what you need to produce: a validated set of enhancer-target gene pairs, a genome-wide regulatory interaction map, or a focused interaction profile at specific loci.

Start with HiChIP when you have a mark or protein of interest and want enrichment for regulatory contacts. H3K27ac HiChIP is the most common configuration for enhancer-target gene mapping and is appropriate for most projects that begin with ATAC-seq or ChIP-seq data identifying active regulatory elements. It is also the most cost-efficient route to a regulatory interaction map, requiring fewer reads and less sequencing depth than genome-wide alternatives. HiChIP sequencing at CD Genomics supports both standard and custom antibody configurations for transcription factors, histone marks, and architectural proteins.

Start with Capture Hi-C when you have a predefined list of regions and need saturated contact coverage at those loci. This is the typical scenario for GWAS follow-up: you have 50–500 fine-mapped variants or candidate enhancers, and you need to know every promoter they contact. The upfront investment in probe design is repaid by the depth of coverage at your regions of interest and the ability to reuse the probe set across multiple samples and conditions. Capture Hi-C provides high-resolution interaction maps at user-defined genomic regions.

Start with Micro-C when you need an unbiased genome-wide map at the highest resolution achievable. This includes projects where enhancer candidates have not been narrowed to a manageable list, where regulatory contacts may involve unannotated or unmarked elements, or where resolving the fine structure of interaction domains — stripes, sub-TAD boundaries, and multi-enhancer hubs — is biologically important. Micro-C XL extends this resolution to mammalian samples with dual-crosslinking chemistry for improved signal retention.

In many projects, these methods are complementary. A common strategy is to perform H3K27ac HiChIP or Micro-C for discovery, then validate and saturate key loci with Capture Hi-C in additional samples or conditions. The discovery phase identifies candidate enhancer-promoter contacts; the focused phase confirms and quantifies them across biological contexts.

Figure 2: Method selection framework — start with the type of prior knowledge (protein/mark, region list, or none) and follow the decision path to the most appropriate method.

Study Design Considerations

Biological replicates. Chromatin interaction data is inherently noisy, and contact frequencies at individual locus pairs can vary substantially between replicates. Two to three biological replicates per condition are recommended; for primary tissue or heterogeneous samples, consider additional replicates to capture biological variability. Pseudo-replicates generated by splitting a single library do not substitute for independent biological replicates and should not be reported as such.

Controls. HiChIP and Capture Hi-C both include internal controls: HiChIP incorporates an input control (non-immunoprecipitated library) to assess enrichment efficiency, and Capture Hi-C includes off-target regions to estimate capture specificity. Micro-C, being unbiased, does not require enrichment controls, but MNase digestion efficiency must be titrated per sample type — under-digestion leaves long chromatin fragments that reduce resolution, while over-digestion depletes nucleosome-protected fragments and biases the library toward linker regions.

Sequencing depth. The depth required depends on the method and the question. For HiChIP, 50–100 million paired-end reads typically provides sufficient coverage for regulatory loop calling at 1–5 kb resolution. For Capture Hi-C, depth is calculated per captured region rather than genome-wide — 100–500 reads per captured restriction fragment is a reasonable target. For Micro-C, 300–500 million reads per sample are recommended for nucleosome-resolution loop calling genome-wide; lower depth (100–200 million reads) can resolve TADs and compartments but may miss individual enhancer-promoter loops.

Integrating orthogonal data. Chromatin interaction data is most interpretable when combined with orthogonal epigenomic annotations. ATAC-seq or DNase-seq identifies accessible regulatory elements within interacting regions. H3K27ac ChIP-seq distinguishes active from poised enhancers and promoters. RNA-seq confirms that the putative target gene is expressed in the relevant cell type. The Fulco et al. ABC model formalizes this integration: enhancer activity (from ATAC-seq and H3K27ac) is weighted by Hi-C contact frequency to produce a quantitative enhancer-gene score that substantially outperforms either data type alone. ATAC-seq and RNA-seq integration analysis provides the expression and accessibility context necessary for interpreting interaction maps.

From Interaction Maps to Validated Regulatory Interactions

A chromatin interaction map identifies candidate enhancer-promoter pairs — regions that are in physical proximity. It does not, by itself, establish that the enhancer regulates the promoter. Validation requires functional perturbation that demonstrates a causal relationship between enhancer activity and target gene expression.

Three validation strategies are commonly applied:

CRISPR interference (CRISPRi) targets a dCas9-KRAB fusion to the candidate enhancer, silencing it, and measures the effect on target gene expression by qPCR or RNA-seq. A decrease in target gene expression upon enhancer silencing, with no effect on neighboring control genes, provides strong evidence for a functional regulatory relationship. The ABC model paper (Fulco et al., 2019) validated its predictions using CRISPRi-FlowFISH at thousands of candidate loci.

Genetic deletion removes the enhancer sequence entirely using paired Cas9 guide RNAs flanking the enhancer. Deletion is more conclusive than silencing — it eliminates the enhancer rather than repressing it — but is lower-throughput and may produce different phenotypes depending on whether the enhancer contributes additively or redundantly with other regulatory elements.

Orthogonal interaction confirmation uses an independent method — such as 3C-qPCR at a specific locus, or DNA FISH visualizing the spatial proximity of the enhancer and promoter in individual nuclei — to confirm that the physical contact observed in the genome-wide interaction map is not a technical artifact of ligation or enrichment.

For projects where enhancer-target gene assignment feeds into target discovery, transcription factor target gene solutions provide integrated workflows from chromatin interaction mapping through functional validation.

Figure 3: From chromatin interaction maps to validated regulatory interactions — discovery, CRISPRi perturbation, and orthogonal confirmation.

Summary

The essential logic of method selection for enhancer-target gene mapping is:

• HiChIP when your question starts with a protein or histone mark — "what do H3K27ac-marked enhancers contact?" — and you want efficient, regulatory-focused interaction maps
• Capture Hi-C when your question starts with a list of loci — "what genes do these GWAS variants or candidate enhancers contact?" — and you need saturated contact coverage at those specific regions
• Micro-C when your question starts without prior assumptions — "what is the complete regulatory interaction landscape of this locus?" — and you need the highest unbiased resolution achievable

In practice, these methods are complementary. An H3K27ac HiChIP experiment identifies candidate enhancer-promoter contacts across the active regulatory genome. A Capture Hi-C follow-up saturates those contacts at top candidate loci. Micro-C resolves the fine structure of multi-enhancer regulatory hubs. The choice is not which method is best in general — it is which method best addresses your specific question with your specific starting material and validation path.

For teams planning enhancer-target gene mapping studies, CD Genomics supports HiChIP, Capture Hi-C, and Micro-C with full study design support, epigenomic data analysis services encompassing interaction calling, enhancer annotation, and target gene assignment, and Hi-C data analysis covering basics through advanced multi-sample comparisons.

Frequently Asked Questions

1. Can I use standard Hi-C instead of these specialized methods for enhancer-target gene assignment?

Standard Hi-C can identify enhancer-promoter contacts, but at typical sequencing depths (200–500 million reads for the human genome), resolution is limited to 5–40 kb bins. This is insufficient to resolve individual enhancer-promoter loops, which often span 10–100 kb and are separated by multiple regulatory elements within a single bin. HiChIP, Capture Hi-C, and Micro-C each address this resolution problem in different ways — through enrichment, capture, or finer fragmentation, respectively — at lower cost than the multi-billion-read Hi-C experiments that would be required for comparable resolution genome-wide.

2. How do I decide between H3K27ac HiChIP and Capture Hi-C for enhancer-target gene mapping?

Choose H3K27ac HiChIP if you want interaction maps centered on active regulatory elements and do not have a predefined, fixed list of loci. Choose Capture Hi-C if you have a specific list of candidate enhancers or variants — particularly from GWAS — and need saturated interaction coverage at those positions. The methods answer related but distinct questions: HiChIP asks "what enhancers and promoters interact in this system?" while Capture Hi-C asks "what does this specific region interact with?"

3. How much starting material does each method require?

HiChIP typically requires approximately 1 million cells per immunoprecipitation, and multiple IPs can be performed from a single crosslinked cell pellet. Capture Hi-C requires 10–25 million cells for the initial Hi-C library, though this library can be used for multiple capture reactions. Micro-C requires 1–5 million cells for mammalian samples with dual-crosslinking. For rare primary cell populations, HiChIP or Micro-C are generally more feasible than Capture Hi-C.

4. Can these methods be applied to frozen or archived samples?

HiChIP and Micro-C work with flash-frozen crosslinked cell pellets and, with protocol optimization, frozen tissue powder. The critical variable is the quality of crosslinking at the time of collection — under-fixed samples cannot be rescued after freezing. Capture Hi-C requires high-molecular-weight DNA from the Hi-C library preparation step, and degradation during sample storage can reduce library quality. Fresh or properly flash-frozen samples processed within 6–12 months of collection produce the most reliable results across all three methods.

5. How do I integrate chromatin interaction data with ATAC-seq and RNA-seq for enhancer validation?

The most widely adopted framework is the Activity-by-Contact (ABC) model (Fulco et al., 2019), which computes a quantitative enhancer-gene score as the product of enhancer activity (geometric mean of ATAC-seq and H3K27ac signal) and Hi-C contact frequency, normalized against all candidate enhancers within 5 Mb of the target gene. The model is implemented in open-source software and can use HiChIP, Capture Hi-C, or Micro-C data as the contact frequency input. Predicted enhancer-gene pairs are then validated by CRISPRi, genetic deletion, or orthogonal biochemical confirmation.

References

1. Mumbach, Maxwell R., Adam J. Rubin, Ryan A. Flynn, Chao Dai, Paul A. Khavari, William J. Greenleaf, and Howard Y. Chang. "HiChIP: efficient and sensitive analysis of protein-directed genome architecture." Nature Methods, vol. 13, no. 11, 2016, pp. 919–922. DOI: 10.1038/nmeth.3999
2. Mifsud, Borbala, Filipe Tavares-Cadete, Alice N. Young, Robert Sugar, Stefan Schoenfelder, et al. "Mapping long-range promoter contacts in human cells with high-resolution capture Hi-C." Nature Genetics, vol. 47, no. 6, 2015, pp. 598–606. DOI: 10.1038/ng.3286
3. Hsieh, Tsung-Han S., Claudia Cattoglio, Elena Slobodyanyuk, Anders S. Hansen, Oliver J. Rando, Robert Tjian, and Xavier Darzacq. "Resolving the 3D Landscape of Transcription-Linked Mammalian Chromatin Folding." Molecular Cell, vol. 78, no. 3, 2020, pp. 539–553.e8. DOI: 10.1016/j.molcel.2020.03.002
4. Fulco, Charles P., Joseph Nasser, Thouis R. Jones, Glen Munson, Drew T. Bergman, et al. "Activity-by-contact model of enhancer-promoter regulation from thousands of CRISPR perturbations." Nature Genetics, vol. 51, no. 12, 2019, pp. 1664–1669. DOI: 10.1038/s41588-019-0538-0

CD Genomics provides epigenomics services for research use only and they are not intended for clinical diagnosis or treatment.