Circular DNA can be found across organisms and cell types. In eukaryotes, researchers often distinguish eccDNA (typically smaller circles) from ecDNA (often larger and structurally complex circles). Because circular DNA can be rare and difficult to capture with standard "linear-first" workflows, study design matters.

In this guide, you'll learn how to:

• Choose a workflow: Circle-seq, long reads, or hybrid designs
• Call circular DNA reliably using breakpoint evidence
• Prioritize candidates for annotation and functional follow-up

New to the topic? Start with eccDNA 101.

eccDNA vs ecDNA roadmap.

Why circular DNA is easy to miss: background, bias, and breakpoints

Most sequencing and analysis pipelines assume DNA is linear. Circular DNA challenges those assumptions in several ways:

• Low signal vs high background: circular molecules can be a small fraction of total DNA, so reads get dominated by linear fragments unless you enrich or design specifically for circles.
• Big size range: small eccDNA (hundreds to thousands of bp) behaves very differently from larger ecDNA (tens of kb to Mb). One workflow rarely performs equally well across sizes.
• Repeat-rich regions and mapping ambiguity: circle detection often relies on breakpoint (junction) evidence (split reads/discordant pairs). Repeats and low mappability can inflate false positives or hide real events.
• Amplification distortion (when used): rolling circle amplification (RCA) can boost sensitivity but may skew representation across circle sizes or sequences.

A practical way to plan is to treat circular DNA studies as two linked tasks: capture (get circle signal into your libraries) and evidence integration (call circles with transparent criteria).

Define your goal first: discovery, quantification, structure

Before selecting a method, decide what "success" looks like for your project. Most avoidable problems come from using a workflow optimized for the wrong goal.

1. Discovery (atlas building)
   Find as many circles as possible and map where they come from.
2. Quantification (comparative studies)
   Compare conditions (e.g., perturbations, timepoints) and interpret differences robustly.
3. Structure (often ecDNA-focused)
   Resolve complex architectures and rearrangements, not just individual breakpoints.
4. Cell-to-cell variation
   Understand heterogeneity across cells or subpopulations (literature reports have described approaches for this purpose).

If you have multiple objectives, pick one primary goal and treat the others as secondary—this makes trade-offs explicit.

How to choose a method: Circle-seq, long reads, and hybrid designs

Below is a decision-first overview of common approaches. Think of it as a planning shortcut rather than a complete catalog.

Method selection overview for eccDNA/ecDNA studies based on your primary goal.

Related reading: Comparative Analysis of eccDNA Detection Methods

Quick decision rules

• Choose short reads when you need throughput and breakpoint-level precision across many samples.
• Choose long reads when structure and full-length resolution are central.
• Choose a hybrid design when you need both structural coherence and precise breakpoint support.
• Choose cell-to-cell variation–focused approaches only when heterogeneity is the question (not as an optional add-on)

How to call circular DNA reliably: evidence, filters, and validation

Calling circular DNA is rarely a one-tool, one-threshold exercise. A reliable analysis plan makes evidence requirements explicit and communicates uncertainty clearly.

Evidence types to combine

1. Junction evidence
   • Split reads spanning the breakpoint junction
   • Discordant read pairs supporting abnormal adjacency
2. Local signal behavior
   • Coverage/enrichment patterns consistent with circularization or the chosen library strategy
   • (For some ecDNA projects) copy number context may be informative, depending on design
3. Structural coherence (especially for long reads)
   • Reads spanning expected circular structures or supporting consistent deconvolution/reconstruction
   • Agreement with breakpoint graphs/paths in reconstruction workflows

Evidence types and confidence tiers for reliable circular DNA calling.

A simple confidence ladder (easy to report)

• Tier 1 (Candidate): at least one junction signal with acceptable mapping quality
• Tier 2 (Supported): multiple independent junction reads and consistent local signal behavior
• Tier 3 (High confidence): reproducible across biological replicates and/or supported by an orthogonal approach
• Tier 4 (Functional lead): high-confidence circles that also align with downstream functional signals

Tool choice: match callers to data type

• Circle-Map is commonly used for short-read breakpoint calling.
• ecc_finder is often used for workflows that may involve short- and/or long-read inputs (depending on study design).
• AmpliconArchitect is widely used for ecDNA/amplicon reconstruction from WGS in research settings.

When feasible, consider a robustness check (e.g., a second caller or a second parameter regime) and define a transparent rule for what you keep.

• For a step-by-step walkthrough from QC to calling and annotation, see our pipeline guide.
• For orthogonal confirmation of prioritized junctions, see validation options.

After detection: annotation, prioritization, and functional follow-up

A circle callset becomes more useful once it is organized into interpretable categories and connected to a clear hypothesis.

A simple path from circular DNA calls to annotation, prioritization, and follow-up analysis.

Step 1: Annotate what each circle contains (or where it comes from)

For each circle (or circle-producing region), annotate:

• gene overlap and gene proximity
• regulatory annotations (promoters/enhancers) when available for your organism/tissue
• repeat content (transposons, satellites, segmental duplications)
• genomic context relevant to your model (e.g., repair- or stress-associated features, if applicable)

Step 2: Prioritize candidates with explicit criteria

A practical rubric often includes:

• confidence tier (from the ladder above)
• recurrence across replicates/samples
• condition-specific change (for comparative studies)
• mappability/repeat risk flags (to avoid fragile calls)
• concordance with other signals (if available)

Step 3: Integrate one complementary dataset (when available)

Pick integration that matches the question:

• Circular DNA + RNA-seq: do nearby genes/pathways show concordant changes?
• Circular DNA + ATAC-seq/ChIP-seq: do circle-producing regions align with chromatin state shifts?
• ecDNA structure + WGS CNV + long reads: do reconstructions explain amplification and rearrangement patterns?
• Circular DNA + cell-state markers (subpopulation-level): do circles enrich in specific subpopulations?

Even one well-chosen integration can move you from "detected" to "interpretable."

If methylation state is relevant to your hypothesis, you may consider adding eccDNA methylation profiling as an orthogonal layer.

Example workflows: bulk eccDNA, ecDNA structure, and cell-to-cell variation mapping

These workflows are research planning templates you can adapt to your organism, sample constraints, and goals.

Three study templates: bulk eccDNA discovery, ecDNA structure reconstruction, and published heterogeneity-focused approaches.

Workflow A: Bulk eccDNA discovery and comparison

1. Choose an enrichment + short-read plan aligned to sample type and input constraints
2. Run a breakpoint-focused short-read caller
3. Apply mapping/QC filters and define tiers (candidate → high confidence)
4. Use replicate behavior to stabilize conclusions
5. Summarize hotspots, recurrent regions, and condition-associated changes
6. Validate a short list of top candidates with an orthogonal approach when feasible

Workflow B: ecDNA structure reconstruction (WGS + reconstruction + long-read support)

1. Generate WGS for CNV and breakpoint landscape
2. Reconstruct amplicon/ecDNA architecture using an appropriate reconstruction tool
3. Use long reads to resolve repeats and support structural paths
4. Connect structure to downstream signals (expression/chromatin) when relevant
5. Report clearly what is directly supported by reads vs inferred by reconstruction

Workflow C: Cell-to-cell variation mapping (heterogeneity-focused approaches)

1. Use a heterogeneity-focused design aligned to your sample type and expected sparsity
2. Identify circle-producing regions and per-cell junction evidence
3. Quantify heterogeneity (presence/absence, frequency classes, cluster enrichment)
4. Integrate with cell-state features when available
5. Prioritize signals that are consistent at the level your design supports (replicates, subpopulations)

FAQ: common choices and mistakes

How much does RCA bias matter?
RCA can increase sensitivity—especially for small circles—but may distort representation. For comparative studies, replication and consistent processing are essential, and conclusions should emphasize reproducible patterns rather than single-sample extremes.

Short reads or long reads—which should I pick?
Short reads are strong for throughput and breakpoint precision; long reads are strong for structural resolution. The right answer depends on whether your primary goal is discovery/quantification or structural reconstruction.

Is one caller enough?
Sometimes, but repeat content, mapping ambiguity, and filtering thresholds can change outcomes. If possible, run a robustness check (second caller or parameter regime) and document your decision rules.

What should I report to make the study reproducible?
Evidence types used (split/discordant/coverage/structure), key thresholds, confidence tiers, replicate behavior, and any orthogonal support for prioritized candidates.

Conclusion: a checklist for a solid circular DNA study

A reliable eccDNA/ecDNA project usually comes down to a few consistent choices:

• Start by naming the primary goal (discovery, quantification, structure, or cell-to-cell variation)
• Choose a workflow that optimizes for that goal, not a generic "best method"
• Define evidence requirements up front (junction + local signal + structure coherence when relevant)
• Communicate results using confidence tiers and transparent filtering rules
• Prioritize candidates systematically (recurrence, robustness, functional context)
• Add one strong functional follow-up step (annotation + a complementary omics layer when available)

With these principles, circular DNA studies become easier to interpret, easier to reproduce, and more likely to generate clear, testable biological conclusions.

CD Genomics circular DNA services

CD Genomics supports circular DNA research projects with the following research-use-only options:

• eccDNA sequencing (e.g., Circle-seq–style enrichment + short-read sequencing)
• Bioinformatics support for circular DNA calling, filtering, and annotation
• Follow-up validation with junction-specific eccDNA qPCR quantification
• Optional add-on: eccDNA methylation sequencing for epigenetic context (when relevant)

These modules can be combined into an end-to-end workflow—from study design to data analysis and follow-up—based on your research goals and sample constraints.
Services listed below are provided for research use only and can be combined based on project goals and sample constraints.