• Home
• Resource
• Decision Guide: Hi-C vs Micro-C vs Capture Hi-C vs HiChIP — How to Choose the Right 3D Genomics Method

TL;DR: Fast method selection in 60 seconds

If you need genome-wide architecture at kb scale, start with Hi-C.

If you need nucleosome-level resolution for fine loops, choose Micro-C.

If you only care about specific loci, use Capture Hi-C.

If your question is protein-centred, pick HiChIP.

Explore service deliverables:

• Genome-wide baseline: Hi-C Sequencing Service
• Nucleosome-scale maps: Micro-C Service
• Targeted loci: Capture Hi-C Sequencing Service
• Protein-anchored loops: HiChIP Service

Hi-C vs Micro-C: the digestion chemistry is the real difference

Direct answer: If your project lives or dies on fine loops, Micro-C's MNase fragmentation is the reason it can out-resolve Hi-C; if you need a robust genome-wide baseline with fewer experimental choke points, Hi-C is often the safer first choice.

Most teams compare Hi-C and Micro-C as if "resolution" were a simple checkbox. In practice, resolution is an outcome—and the most important upstream driver is how chromatin gets fragmented before proximity ligation. That one choice shapes fragment size distributions, sequence bias, and how evenly contacts populate across the genome. It also shapes whether your downstream loop or domain calls will be stable across replicates.

So the decision is not "which method is better." The decision is "which fragmentation chemistry fits my hypothesis and my constraints." If your sample is scarce, or your budget cannot support deep sequencing, a method that looks superior on paper can still be the wrong choice.

Micro-C uses MNase digestion for nucleosome-scale fragments, while Hi-C relies on restriction enzymes, shaping resolution and loop detection.

MNase digestion (Micro-C) vs restriction enzymes (Hi-C): what changes in practice?

Direct answer: Restriction enzymes cut at sequence motifs and introduce motif-driven coverage unevenness, while MNase digests linker DNA between nucleosomes and can produce more uniform, nucleosome-sized fragments.

Hi-C workflows usually digest crosslinked chromatin with a restriction enzyme (for example, enzymes cutting at specific recognition motifs). Because motifs are not uniformly distributed, your fragments are not uniformly sized, and some regions naturally receive fewer informative ligation junctions. This is not "bad" by default—Hi-C pipelines and interpretation norms were built around it—but it is a real source of bias when you push toward finer bins.

Micro-C replaces restriction digestion with micrococcal nuclease (MNase), which trims linker DNA between nucleosomes. In well-controlled digestions, this generates predominantly mono-nucleosome fragments and a tighter size range. That is why Micro-C can support nucleosome-level contact maps and sharpen short-range features, especially loop-like peaks and stripes at smaller genomic distances.

The trade-off is control. MNase has a narrower "golden window." Under-digestion leaves long fragments and blurs fine structure. Over-digestion can reduce informative junctions and destabilise library complexity. That means Micro-C can demand more aggressive QC at the digestion and size-selection stages, particularly when sample quality varies.

Micro-C was originally demonstrated at nucleosome resolution in yeast, where it revealed fine self-associating domains that were hard to see at coarser scale. (Hsieh et al., 2015. DOI: https://doi.org/10.1016/j.cell.2015.05.048)

Resolution expectations: kb-level vs nucleosome-level (what you actually get)

Direct answer: Resolution is not a promise; it scales with usable unique contacts, library complexity, and sequencing depth—so plan depth around your bin size and deliverable.

When someone says "kb resolution," ask a follow-up question: "At what confidence, in what cell type, with what depth?" In contact mapping, resolution is essentially the smallest bin size that still contains enough unique contacts to support stable features after filtering, deduplication, and normalisation. The number of reads is only part of the story; library complexity and the proportion of informative cis contacts often matter as much.

Hi-C can achieve very high resolution, but it typically requires substantial sequencing depth and careful processing for dense maps. A classic in situ Hi-C dataset demonstrated kilobase-scale mapping in human cells, but it relied on very deep coverage. (Rao et al., 2014. DOI: https://doi.org/10.1016/j.cell.2014.11.021)

Micro-C can generate nucleosome-scale contact patterns and often improves the contrast of short-range features. In mammalian cells, Micro-C has been shown to reveal many additional looping interactions compared with Hi-C, especially at shorter distances where nucleosome-scale fragmentation helps populate fine bins. (Krietenstein et al., 2020. DOI: https://doi.org/10.1016/j.molcel.2020.03.003)

A practical way to set expectations is to decide what you need to call:

• If your deliverable is compartments and broad domains, Hi-C is often sufficient.
• If your deliverable is fine loops and short-range structure, Micro-C is often worth it.
• If your deliverable is a shortlist of loci, consider concentrating depth with Capture Hi-C instead of scaling whole-genome depth.

TADs and loops: which method is more precise, and why?

Direct answer: Hi-C is strong for large-scale organisation, while Micro-C often improves fine loop detection by sharpening short-range contact contrast.

TADs and compartments are relatively robust, because they reflect broad patterns that remain visible even when you bin coarsely. Hi-C tends to perform well here, and many comparative studies, tools, and community benchmarks are built around Hi-C-like data structures.

Loops are different. Loop peaks can be narrow and local, and they compete against background contact decay. Micro-C's nucleosome-scale fragmentation can make loop peaks stand out more clearly, especially when your question is driven by enhancer–promoter wiring or fine architectural changes. That said, more loops in a call set is not automatically better; it can also reflect differences in depth, filtering, and calling thresholds.

For decision-making, focus on reproducibility rather than raw counts:

• Are loop calls consistent across biological replicates?
• Do contact maps show stable decay curves and low noise?
• Do key features persist under reasonable binning choices?

Sample sensitivity: which is more likely to fail with precious material?

Direct answer: Micro-C can be less forgiving because MNase digestion and chromatin integrity are tighter bottlenecks, while Hi-C is often more tolerant of variability.

Both methods require good nuclei and controlled crosslinking. However, Micro-C typically adds sensitivity at the digestion stage, because MNase performance depends strongly on chromatin accessibility, nuclease activity, and timing. If you have irreplaceable samples, the safest "first pass" is often the method with fewer choke points and more mature QC norms.

If your project is high-stakes, consider a conservative strategy:

• Use Hi-C as a robust baseline to validate architecture changes.
• Use Micro-C selectively when fine loops are essential deliverables.
• Use targeted methods when your hypothesis is locus-limited.

Next step: If you are choosing between Hi-C and Micro-C for a specific project, the fastest de-risking move is aligning on deliverables + QC gates upfront. You can compare package options on our Hi-C Sequencing Service and Micro-C Service pages.

When to choose Micro-C (and when not to)

Direct answer: Choose Micro-C when your success metric depends on nucleosome-level detail and fine loop discovery; skip it when sample quality is uncertain, or when your decision can be made at kb-scale.

Micro-C is not a "better Hi-C."

It is a different trade: you buy finer structure, but you accept tighter experimental tolerances. If your project has a hard deadline, or if your material is limited and heterogeneous, that trade deserves a sober look.

Best-fit scenarios for Micro-C

Direct answer: Micro-C is most valuable when you need to resolve short-range contacts that Hi-C tends to blur.

Micro-C is a strong choice when:

• Fine enhancer–promoter wiring is central to your hypothesis.
• You expect many closely spaced loops within active regulatory neighborhoods.
• You need nucleosome-position context alongside 3D contacts.
• You plan to compare subtle structural differences across conditions.

In mammalian cells, ultra-deep Micro-C maps captured known higher-order features while improving short-range signal and loop discovery. (Krietenstein et al., 2020. DOI: https://doi.org/10.1016/j.molcel.2020.03.003)

Red flags: when Micro-C is the wrong tool

Direct answer: If the project can succeed with domain-level calls, Micro-C may add cost and risk without adding actionable insight.

Micro-C is often a poor fit when:

• Your sample quality varies, or nuclei prep is inconsistent.
• You cannot support the depth needed for your bin size goals.
• Your deliverable is compartments, TADs, or coarse domain shifts.
• You only care about a defined gene set or a small locus list.

If your question is locus-limited, a targeted method often gives a clearer ROI than pushing Micro-C depth. That is a "selection" decision, not a prestige decision.

What to demand from any Micro-C provider

Direct answer: Ask for digestion control evidence and decision-ready QC, not just a final heatmap.

Request QC elements that de-risk the project:

• MNase digestion optimisation and fragment size profiles.
• Library complexity and duplicate rates after filtering.
• Cis/trans balance and distance-dependent decay behaviour.
• Replicate concordance metrics and clear failure criteria.

If a provider cannot describe what would trigger a rework, you are absorbing hidden risk.

When to choose Hi-C (the baseline decision that saves budgets)

Direct answer: Choose Hi-C when you need a robust, interpretable genome-wide baseline for architecture and comparisons, especially when you value mature analysis norms.

Hi-C remains the most widely used starting point for 3D genome studies for one reason: it often gives you stable "big-picture" structure with fewer experimental choke points. For many teams, that is the most rational first decision, particularly when the goal is to screen conditions, map broad domain changes, or generate a publishable baseline before moving to more specialised assays.

Best-fit scenarios for Hi-C

Direct answer: Hi-C is a strong default for broad architecture and hypothesis generation.

Hi-C fits well when:

• You want compartments, contact domains, and global organisation.
• You need a cross-condition comparison framework with standard tools.
• You are unsure whether the effect is global or local.
• You want a baseline dataset to justify a deeper second assay.

A widely cited in situ Hi-C study produced kilobase-scale maps in human cells with very deep sequencing and careful processing, illustrating how far Hi-C can go when depth is available. (Rao et al., 2014. DOI: https://doi.org/10.1016/j.cell.2014.11.021)

What Hi-C does better than Micro-C in practice

Direct answer: Hi-C typically tolerates more variability in input and digestion, making it a safer baseline when constraints are tight.

Hi-C advantages often show up as operational stability:

• Less sensitivity to over/under digestion extremes.
• More predictable downstream interpretation for macro-features.
• Mature benchmarking expectations for domains and compartments.
• Easier cross-study comparisons across public datasets.

This matters if your team is using 3D genomics to drive go/no-go decisions. A method that delivers a consistent baseline can be more valuable than one that sometimes delivers sharper detail.

A "minimum viable" Hi-C design mindset

Direct answer: Plan Hi-C around replicates and success criteria, not around a single target bin size.

For selection-focused projects, prioritise:

• Biological replicates first, then depth.
• Clear acceptance metrics (complexity, concordance, noise).
• A binning plan tied to your intended calls (domains vs loops).

If your team cannot articulate what "good enough" looks like, the dataset will be hard to use, even if it is large.

Capture Hi-C: the best choice when your hypothesis is locus-limited

Direct answer: Choose Capture Hi-C when you need high confidence at specific promoters or regions, and you do not want to pay whole-genome sequencing costs to get it.

Capture Hi-C is often the most "budget honest" choice. It admits a simple truth: many projects are not asking genome-wide questions. They are asking region-defined questions, often around promoters, GWAS loci, or curated enhancer sets. In that setting, concentrating sequencing where you need resolution can beat pushing depth on genome-wide assays.

What "targeted" really means: capture design drives everything

Direct answer: Capture Hi-C performance is dominated by probe design, not by the sequencing instrument.

Capture Hi-C uses hybridisation baits to enrich ligation products involving selected fragments. That means the design defines:

• Which loci you can interpret confidently.
• Where "blind spots" will exist by construction.
• How comparable results are across samples and batches.

A classic Capture Hi-C paper mapped long-range promoter contacts across ~22,000 promoters in human blood cell types, highlighting how capture enrichment makes promoter-centred questions scalable. (Mifsud et al., 2015. DOI: https://doi.org/10.1038/ng.3286)

When Capture Hi-C beats deeper Hi-C

Direct answer: If your question is restricted to known loci, capture enrichment can deliver higher effective resolution per dollar.

Capture Hi-C often wins when:

• You have a defined gene list or promoter set.
• You need statistical power around those loci across conditions.
• You want interpretable contacts without a "billions of reads" plan.

A practical comparison question to ask is:

Would you rather distribute reads across 3 billion bases, or concentrate them where you will make decisions?

Common pitfalls that waste projects

Direct answer: Most Capture Hi-C failures come from mismatched design to hypothesis, not from sequencing.

Watch for these pitfalls:

• Panels that miss key regulatory fragments or promoter isoforms.
• Inconsistent designs across cohorts, blocking clean comparisons.
• Over-interpreting weak contacts without replicate support.
• Mixing capture designs and expecting uniform sensitivity.

If your project is selection-driven, design should be treated as part of the experimental method, not as a procurement detail.

A literature-backed "case pattern" for promoter capture

Direct answer: Promoter capture datasets can reveal cell-type specific promoter interaction networks that track lineage and regulatory state.

In a large promoter capture Hi-C study across primary hematopoietic cell types, promoter interactions were highly cell-type specific and enriched for links between active promoters and enhancers. (Javierre et al., 2016. DOI: https://doi.org/10.1016/j.cell.2016.09.037)

For non-clinical research teams, the actionable takeaway is methodological: promoter-focused capture can turn a broad regulatory hypothesis into a testable promoter interaction atlas with manageable sequencing budgets.

HiChIP: the go-to method for protein-centred 3D regulation

Direct answer: Choose HiChIP when your central question is "which contacts are anchored by this protein or histone mark," and you can meet antibody and ChIP-quality requirements.

HiChIP is built for a specific decision context: you are not trying to measure "all contacts." You want the subset of contacts most likely to matter for a protein-defined regulatory mechanism. That makes HiChIP a powerful selection tool when the protein anchor is the real hypothesis driver.

What HiChIP captures that Hi-C and Micro-C do not

Direct answer: HiChIP enriches contact pairs associated with a protein mark, increasing signal-to-noise for that protein's interaction landscape.

In the original HiChIP paper, the method was introduced as a protein-centric chromatin conformation approach that improves the yield of conformation-informative reads and reduces input requirements compared to older ChIA-PET-style strategies. (Mumbach et al., 2016. DOI: https://doi.org/10.1038/nmeth.3999)

That matters for decision-making because you are paying to sequence information you will actually interpret, rather than paying to discover that your protein-relevant contacts are sparse in a genome-wide library.

Antibody and ChIP quality: the hidden make-or-break factor

Direct answer: HiChIP is only as reliable as your antibody specificity and ChIP-grade chromatin.

This is the selection trap many teams miss.

With Hi-C or Micro-C, bias often comes from fragmentation and depth.

With HiChIP, a major bias source is antibody and enrichment performance.

Before committing, align on:

• Antibody validation evidence in your sample type.
• Enrichment QC expectations (e.g., peak quality and reproducibility).
• A plan for what happens if enrichment underperforms.

If you do not have ChIP-grade material, HiChIP may be a high-risk choice even if the scientific logic is perfect.

HiChIP vs Capture Hi-C for "loop" questions

Direct answer: Capture Hi-C is region-driven, while HiChIP is protein-driven; pick the one that matches what you can defend in a paper.

Use Capture Hi-C if:

• Your question is restricted to a gene set or locus list.
• You want consistent sensitivity across those loci.

Use HiChIP if:

• Your hypothesis centres on a protein mark or TF.
• You care most about contacts likely mediated by that factor.

A helpful framing is:

Capture Hi-C answers "where do these loci contact?"

HiChIP answers "where does this factor anchor contacts?"

Deliverables and QC you should request

Direct answer: Demand reproducible, factor-anchored contact calls with clear thresholds.

Ask for:

• Replicate concordance of interaction calls.
• Clear filtering and peak/anchor definitions.
• A description of how loops were called and controlled.
• Contact maps and summary metrics that support interpretation.

Capture Hi-C concentrates sequencing on selected loci; HiChIP concentrates signal on protein-bound contacts—two different paths to higher signal-to-noise.

Input material and sample risk: choose the method that won't waste your samples

Direct answer: If samples are irreplaceable, choose the method with the highest probability of producing interpretable data under your constraints, even if it is not the "highest resolution" method.

Every method comparison should include a risk lens.

Not just "what could I learn," but "what could fail."

This is especially true for primary material and hard-to-replace models.

Practical input tiers: what changes as samples get smaller

Direct answer: Lower input amplifies every bottleneck—complexity loss, noise inflation, and replicate instability.

As input drops, three things usually become harder:

• Maintaining library complexity after filtering.
• Keeping replicate concordance at your intended bin size.
• Supporting confident loop calls without overcalling.

This is why "method choice" and "depth planning" are inseparable. If input is limited, the most expensive mistake is choosing a method that requires perfect execution to succeed.

Where projects fail most often

Direct answer: Most failures trace to digestion control, low complexity, or discordant replicates.

Common failure patterns include:

• Digestion outside the optimal window (both directions).
• Low unique ligation pairs after filtering and deduplication.
• High noise at the distances you care about.
• Replicates that disagree in the features you intend to claim.

A good provider should be able to state failure criteria clearly.

That clarity is part of risk management.

A pre-flight checklist that prevents wasted sequencing

Direct answer: Run a design review that links hypothesis → method → QC gates → success metrics.

Before you start, align on:

• What feature(s) define success (domains, loops, specific loci).
• Replicate plan and minimum acceptance criteria.
• Target depth range tied to deliverables, not vanity bins.
• A rework policy triggered by objective QC gates.

This is where "selection" becomes operational.

Decision matrix: the shortest path to the right choice

Direct answer: Choose by constraint: question type → required resolution → targeted vs genome-wide → protein dependence → sample risk; if two methods fit, pick the one with fewer failure points.

Here is a compact decision logic you can apply quickly.

If you need unbiased discovery

Direct answer: Start genome-wide, then specialise only if needed.

Pick Hi-C when:

• You need compartments or broad domains.
• You need a stable baseline for comparisons.

Pick Micro-C when:

• Fine loops and short-range features are the decision output.
• You have the QC control and depth budget to support it.

If you only need selected loci

Direct answer: Targeted enrichment often beats brute-force sequencing.

Pick Capture Hi-C when:

• You have a defined promoter set or region list.
• You want stronger statistical power at those loci.

If the protein is the anchor

Direct answer: Protein-centric assays are best when the protein defines interpretability.

Pick HiChIP when:

• Your hypothesis is factor-driven.
• Antibody performance is validated and reproducible.

If your aim is publication-grade fine loops

Direct answer: Micro-C can be the best tool when you can support its demands.

Pick Micro-C when:

• Nucleosome-scale mapping is central to your claims.
• Replicates and depth are planned to defend loop calls.

If you are still torn between two methods, choose the one that lets you define success most cleanly, with the fewest assumptions.

A practical decision flowchart for selecting Hi-C, Micro-C, Capture Hi-C, or HiChIP based on question type, sample risk, and constraints.

Next step: turn a method choice into a quote-ready project plan

Direct answer: Once you choose a method, de-risk the project by locking deliverables, QC gates, replicates, and depth planning before sequencing starts.

At this point, you are no longer choosing a method in the abstract.

You are choosing an execution plan that will survive review, revision, and reuse.

To move fast without wasting samples, prepare:

• Your sample type, count, and constraints (quantity and quality).
• Your decision output (domains, loops, loci, or factor-anchored contacts).
• Your comparison design (conditions, replicates, batch structure).
• Your minimum acceptance criteria for QC and concordance.

A good consultation should translate those items into a concrete plan: workflow choice, sequencing strategy, and analysis deliverables aligned to your hypothesis.

FAQ: common decision questions researchers ask

Is Micro-C always better than Hi-C?

Direct answer: No—Micro-C is better for fine loops and nucleosome-scale structure, but Hi-C is often safer for robust baseline architecture under tight constraints.

Micro-C can reveal more short-range features, but it also demands tighter digestion control and often deeper depth planning. If your decision is domain-scale, Hi-C can be the better choice operationally and analytically.

Can Capture Hi-C replace Hi-C for most projects?

Direct answer: Only when your question is truly locus-limited; Capture Hi-C is not designed for unbiased genome-wide discovery.

Capture Hi-C concentrates sequencing on targeted fragments, which is ideal for promoter- or locus-centred hypotheses. If you need global compartments and unbiased architectural shifts, genome-wide assays remain the appropriate baseline.

HiChIP vs Capture Hi-C: which is better for enhancer–promoter loops?

Direct answer: Choose Capture Hi-C when loci define your hypothesis, and HiChIP when a protein mark defines interpretability.

If you need promoter-centric maps regardless of factor binding, capture is often more uniform. If you need loops specifically associated with a mark or TF, HiChIP provides a more direct mechanistic anchor—assuming antibody performance is strong. (Mumbach et al., 2016. DOI: https://doi.org/10.1038/nmeth.3999)

How much sequencing depth do I need for reliable loops or TADs?

Direct answer: Depth depends on the feature you must call and the bin size you need to defend; plan around reproducibility, not a single universal number.

Broad domains and compartments generally stabilise at lower effective resolution. Fine loop detection is more depth-sensitive and more sensitive to noise and complexity. Use replicates and explicit acceptance criteria to avoid over-interpreting weak, non-reproducible calls.