Choosing eccDNA Enrichment Methods: Exonuclease Digestion, RCA, Capture, and Controls

choosing-eccdna-enrichment-methods-exonuclease-digestion-rca-capture-and-controls-1a

Selecting the right eccDNA enrichment strategy is, at its core, a trade-off between specificity and yield. Exonuclease-based cleanups can deliver tight specificity by removing linear DNA; rolling circle amplification (RCA) boosts sensitivity, especially for small circles, at the cost of amplification bias; hybrid capture narrows the scope to known targets with strong signal-to-noise; and ATAC-derived approaches reuse existing data with algorithmic calls. The most important organizing principle for choosing among these eccDNA enrichment methods is your research goal: Discovery versus Targeted.

If you're running discovery—profiling a broad, often unknown repertoire of eccDNA across loci and sizes—you'll generally tolerate some bias to gain sensitivity and breadth. If your goal is targeted validation or quantification of known loci (e.g., oncogene-associated circles), you'll favor methods that deliver specificity, predictable turnaround, and clean, publishable controls.

This guide complements our general Experimental Workflow for eccDNA sequencing, focusing on method selection and practical constraints. For a step-by-step workflow overview (library prep, enrichment, pitfalls), see the Experimental Workflow article in this series: Experimental Workflow for eccDNA Sequencing: Enrichment, Library Prep, and Common Pitfalls.

Quality metrics and QC for eccDNA studies — library preparation checkpoints and run‑readiness in NGS Library Preparation Quality Control.

Core eccDNA Enrichment Methods: How to Choose for Discovery vs Targeted

Before diving into individual protocols, it helps to anchor decision-making in the Discovery vs Targeted dimension and the realities of sample type and DNA integrity.

Discovery: Maximize sensitivity and breadth. Accept some size bias and plan orthogonal validation. Typical stack: exonuclease cleanup + RCA, followed by short-read sequencing and computational filtering; long-read confirmation for large circles.
Targeted: Maximize specificity and interpretability for known loci. Favor hybrid capture and amplicon workflows on exonuclease-cleaned libraries; avoid RCA where possible to minimize bias.

Think of eccDNA enrichment like tuning a radio: RCA amplifies everything in range with a boost to high-frequency signals (small circles), while hybrid capture dials to a specific station (known loci) and keeps noise down. Exonuclease cleanup is your noise filter before any amplification.

Method 1: Exonuclease Digestion (The Standard)

Exonuclease digestion is the foundational cleanup used across eccDNA protocols to remove linear DNA and enrich circular molecules prior to amplification or library preparation. Mechanistically, enzymes such as Exonuclease V (RecBCD) degrade linear duplex DNA in both directions, leaving intact circular templates that lack free ends. Pairing ExoV with ATP-dependent DNases (e.g., Plasmid-Safe) further reduces residual linear contaminants.

How it works (mechanism)

Exonuclease V targets linear dsDNA and ssDNA ends, digesting them processively. Circular DNA forms closed structures without accessible ends, rendering them refractory to ExoV digestion under standard conditions. Many labs optionally add Plasmid-Safe ATP-dependent DNase as a follow-on step to degrade any lingering linear fragments after ligation or partial circularization workflows. These steps collectively improve the circular-to-linear ratio before downstream analysis or amplification.

Mechanistic details and reagent properties are documented in enzyme vendor literature and methods reviews, e.g., NEB's Exonuclease V product information and methods comparisons. See the general overview in comparative analyses of eccDNA methodologies (2024) and NEB's ExoV technical documentation: NEB Exonuclease V (RecBCD) product page.

When to use it

Discovery goals: Use as a cleanup step prior to RCA to reduce linear templates and increase the proportion of eccDNA-derived reads.
Targeted goals: Use prior to hybrid capture or amplicon workflows to cut background and improve on-target capture efficiency.
Sample types: Works best with high molecular weight (HMW) genomic DNA (high-integrity cell lines). Fragmented inputs (FFPE, small biopsies) may still benefit but expect lower yield; adjust digestion times and units cautiously.

Micro-protocol (lab-dependent; pilot titration recommended)

Input DNA: 0.5–2 µg HMW genomic DNA in the vendor buffer.
Add ExoV per vendor units/µg (follow kit guidance); incubate 37°C for 30–60 minutes.
Optional: Add Plasmid-Safe ATP-dependent DNase with ATP; incubate 30–60 minutes.
Cleanup: Perform magnetic bead purification (1.3×–1.8× SPRI ratio depending on desired size selection). Elute in 10 mM Tris-HCl, pH 8.5.
Optional second cleanup: If inhibitors persist, repeat bead clean or perform a column cleanup.

For general library preparation considerations (fragment size distributions, molarity, QC checkpoints), see NGS Library Preparation Quality Control.

Controls and QC

Negative control: A mock digestion without enzyme to estimate background.
Positive control: Include a small plasmid or circularized amplicon spike-in at known copy number to quantify digestion selectivity.
QC metrics: qPCR/dPCR to measure circular-to-linear ratio; fragment analysis (Femto Pulse) to confirm integrity; replicate concordance for called circles.
Acceptance thresholds (lab-dependent): ≥3× increase in circular-to-linear ratio post-digestion; <20% reads aligning to canonical linear fragments in the enriched library.

Pros and cons

Pros:

High specificity for removing linear DNA; improves downstream signal-to-noise.
Flexible precursor to RCA, hybrid capture, or amplicon validation.

Cons:

Requires HMW input for best performance; fragmented samples produce lower circular recovery.
Enzyme cost and lot-to-lot variability may necessitate titration.

Troubleshooting checklist

Over-digestion reducing circular yield? Reduce enzyme units or incubation time; verify buffer conditions.
Residual linear DNA? Add a PSAD step or repeat ExoV with adjusted units; verify ATP source for PSAD.
Low recovery from FFPE? Consider shorter digestions, adjust bead ratios to favor larger fragments, or pivot to targeted capture for validation.
Inconsistent replicates? Standardize input integrity checks and run paired mock digestions to benchmark variability.

Method 2: Rolling Circle Amplification (RCA) with Phi29

RCA is the workhorse for discovery-oriented profiling. Using Phi29 DNA polymerase, RCA amplifies circular DNA isothermally by strand displacement, generating long concatemers that translate into abundant reads after fragmentation and library prep.

Secondary keyword: rolling circle amplification eccDNA

How it works (mechanism)

Phi29 polymerase exhibits strong strand displacement and high processivity, making it ideal for amplifying circular templates. Priming is typically performed with random hexamers, sometimes modified to resist exonucleases. The result is a rolling replication around the circle, producing long concatemers comprising repeating units of the circular template.

Method papers document the widespread use of RCA in Circle-Seq-derived workflows; for a broad survey of performance and limitations, see the 2024 comparative analysis: Comparative analysis of eccDNA methodologies, and a representative Circle-Seq application: Genome-wide characterization of eccDNA (2023).

Known biases and why they matter

Size bias: RCA preferentially amplifies small circles (often <2–10 kb), causing over-representation of these species relative to larger circles.
Coverage artifacts: Libraries derived from RCA often show sharp coverage spikes and periodic patterns corresponding to concatemer repeats.
Quantification caveats: Copy number estimates based on RCA-amplified libraries can misrepresent true abundance; favor relative, not absolute quantitation.

Mitigation strategies include shortening amplification times (e.g., 2–4 hours), tuning primer chemistry, pairing RCA with prior exonuclease cleanup, and applying bioinformatics filters to remove concatemer-induced artifacts. For isothermal amplification guidance and kit documentation, consult vendor resources such as NEB phi29-XT RCA kit.

Micro-protocol (lab-dependent; pilot titration recommended)

Input DNA: 10–100 ng of circular-enriched DNA.
Priming: Exonuclease-resistant random hexamers (where available) at vendor-recommended concentrations; optional pyrophosphatase to suppress dNTP depletion.
Amplification: 30°C for 2–4 hours (optimize by input and observed bias). Avoid over-extension which increases artifact risk.
Post-RCA processing: Shear to 200–500 bp (enzymatic or sonication) and proceed to Illumina-compatible library prep.
Optional: Include a brief exonuclease step post-RCA if linearized products contaminate, but be cautious not to degrade genuine circular concatemer-derived products.

Controls and QC

Spike-in circles of defined sizes (e.g., 300 bp, 1 kb, 3 kb) to quantify size bias.
Negative controls without RCA to estimate background.
Coverage-based QC: Assess spike-like coverage and duplication rates; ensure replicate concordance.
Acceptance thresholds (lab-dependent): Duplication rate within expected range for RCA libraries; detectable recovery of spike-in circles across sizes with known skew favoring smaller ones.

Pros and cons

Pros:

Excellent sensitivity for low-input discovery; particularly strong for small circles.
Simple isothermal workflow; widely available reagents.

Cons:

Favoring smaller circles can skew profiles; quantitation requires caution.
Concatemer artifacts complicate coverage interpretation and downstream analysis.

Troubleshooting checklist

Excessive coverage spikes? Shorten RCA time, adjust primer mix, reduce input complexity with prior exonuclease.
Poor amplification from very low input? Verify primer quality and polymerase activity; consider pre-cleanup to remove inhibitors.
Over-representation of tiny circles? Incorporate spike-ins with known sizes to quantify and later correct bias in analysis.
Library prep failures post-RCA? Ensure proper shearing and cleanup, as viscous concatemer products can impede adapter ligation.

For method background and protocol context, see CD Genomics' main service pages for practical support and confirmation workflows: Genomics Services and Long‑read Sequencing.

Method 3: Hybrid Capture and ATAC-derived Approaches

Hybrid capture and ATAC-derived workflows primarily serve targeted goals—confirming or quantifying suspected eccDNA/ecDNA at known loci. Hybrid capture uses biotinylated probes to enrich for sequences of interest; ATAC-derived detection leverages accessibility-based libraries and algorithmic calling.

Hybrid capture for targeted panels

Hybrid capture can be run on libraries prepared from exonuclease-cleaned DNA to improve on-target rates and suppress background. For known oncogene loci (e.g., EGFR, MYC), custom probe panels can target expected regions and, where feasible, junction-spanning sequences. Because eccDNA/ecDNA can be structurally heterogeneous, panel design benefits from redundancy and includes flanking regions.

Disclosure: CD Genomics is our product. As a neutral, practical example, you can operationalize targeted panels using CD Genomics resources without altering your scientific design. For background on panel design and capture workflows, see Overview of targeted sequencing, and for implementation options, see Human & Mouse Exome Sequencing. These services can support custom probe designs for suspected eccDNA loci and standard capture workflows compatible with RUO studies.

For amplicon-based validation, see Amplicon Sequencing Services. For long-read structural confirmation, see Long-read Sequencing and Nanopore Targeted Sequencing.

ATAC-derived detection (Circle-ATAC concepts)

ATAC-seq-based detection of eccDNA has been demonstrated by reanalyzing accessibility libraries for circular junction signatures. This avoids RCA and can be attractive when you already have deep ATAC datasets. Methods papers have described inverse PCR validation and FISH confirmation from ATAC-derived calls. For an accessible review of categorizing eccDNA across datasets and approaches, see Categorizing eccDNA across human tissues (2024); for a foundational demonstration of ATAC-derived detection, see ATAC-Seq–based identification of eccDNA (2021).

Micro-protocol pointers (hybrid capture; lab-dependent)

Library source: Prefer exonuclease-cleaned DNA. If input is limited, consider low-input capture kits and avoid RCA to reduce bias.
Probe design: Include regions around suspected junctions; add redundancy for heterogeneous breakpoints.
Sequencing depth: 5–20 million paired-end reads per sample depending on panel size; evaluate on-target fraction and uniformity.
Validation: Junction PCR/Sanger across predicted breakpoints; long-read nanopore for structural confirmation when panels suggest complex architecture.

Pros and cons

Pros:

High specificity for known targets; efficient use of sequencing budget.
Easier quantification at targeted loci with clean controls.

Cons:

Risks missing unexpected circles outside the panel.
Junction heterogeneity can challenge probe design; requires iteration.

Validation and Controls

Regardless of method, publication-grade eccDNA studies hinge on robust controls and computational validation. This section outlines practical designs you can adopt and points you to resources covering analysis details.

Synthetic spike-in circles

Design synthetic circular spike-ins to quantify enrichment efficiency and bias. Choose sequences absent from the host genome to avoid mapping ambiguity, and construct circles at multiple sizes (e.g., ~300 bp, ~1 kb, ~3 kb). Add a fixed amount to samples pre-enrichment and measure recovery post-enrichment via qPCR/dPCR and sequencing read counts. Because explicit peer-reviewed protocols vary, parameters are lab-dependent; pilot titration is advised.

Junction PCR and sequencing confirmation

Validate predicted eccDNA with inverse PCR across junctions, followed by Sanger sequencing of the amplicon to confirm the breakpoint. For larger circles or suspected ecDNA structures (>10 kb), perform targeted long-read sequencing to verify circle continuity and map complex architectures.

Computational validation and artifact filtering

Use established pipelines to detect circular junctions and filter artifacts typical of RCA-based libraries. Pipelines such as Circle-Map, ECCsplorer, and recent eccDNA-pipe integrate split-read detection, coverage continuity checks, and scoring thresholds. A 2024 comparative analysis provides benchmarks for method sensitivities and common artifacts. For a dedicated summary of computational steps and reporting standards, see our series article: Bioinformatics for eccDNA: Detection Algorithms, Filtering Artifacts, and Reporting Standards.

QC metrics you should track

Enrichment efficiency: Ratio of circular to linear spike-in recovery pre/post enrichment.
Background fraction: Proportion of reads mapping to linear genomic regions in enriched libraries.
Reproducibility: Overlap of called eccDNA between technical replicates.
Coverage patterns: Presence of RCA-induced spikes vs uniform coverage in exonuclease-only workflows.
Sequencing depth: For discovery, plan 50–100M paired-end reads per sample (lab-dependent). For targeted panels, 5–20M reads depending on panel size.

For practical QC checkpoints during library prep and capture, see NGS Library Preparation Quality Control.

Decision Matrix: Which to Choose?

The decision starts with your goal—Discovery vs Targeted—and is shaped by sample type and DNA integrity. Below is a compact matrix to guide initial selection.

Goal	Sample Type	Recommended Method	Input Requirement	Expected Bias	Turnaround Considerations
Discovery	Cell line, HMW DNA	Exonuclease digestion + RCA	0.5–2 µg DNA → 10–100 ng post-cleanup	Favors small circles (<2–10 kb)	Moderate; RCA adds hours, library prep standard
Discovery	Biopsy/FFPE (low input)	RCA (shorter time) ± exonuclease cleanup	10–50 ng	Strong small-circle bias; fragmentation effects	Moderate; low-input workflows required
Targeted	Known oncogene loci (e.g., MYC)	Hybrid capture on exonuclease-cleaned libraries	≥50 ng (panel-dependent)	Minimal if panel designed well	Efficient; capture adds 1–2 days
Targeted	Rapid validation/quantification	Amplicon + junction PCR/Sanger; optional long-read	10–50 ng	Minimal method bias; locus-specific	Fast; days not weeks

For quality metrics to include in your study design and vendor evaluations, see our series article: Quality Metrics for eccDNA Sequencing: Enrichment Efficiency, Background, and Reproducibility.

To design stress-induced experiments (e.g., hydroxyurea, cell cycle perturbations) and understand method implications, see: Replication Stress and eccDNA: Hydroxyurea, Cell Cycle Effects, and Study Design Considerations.

Figures & Practical Notes

Below are three practitioner-focused figures to visualize core differences across methods and highlight practical equipment.

Diagram contrasting Exonuclease digestion (removing linear DNA) versus Phi29 rolling circle amplification producing concatemers from eccDNA. Figure 1: ExoV/Plasmid-Safe remove linear DNA, leaving circular templates intact. Phi29 RCA amplifies circles into concatemers, increasing yield but introducing small-circle bias.

Simulated coverage plots showing RCA-induced spikes versus smoother coverage after exonuclease cleanup. Figure 2: RCA libraries often show sharp coverage spikes (periodicity) at small-circle loci, while exonuclease-cleaned, non-RCA libraries exhibit more uniform coverage suitable for targeted quantification.

Diagram of magnetic bead cleanup on a rack showing bead pellet, supernatant, washes, and typical bead ratios. Figure 3: Magnetic bead separation stand used for cleanup. Adjust bead-to-sample ratios (1.0×–1.8×) to select for desired fragment sizes and remove enzymes/inhibitors.

Expanded Practical Guidance: Case Scenarios and Tips

To translate the decision matrix into lab actions, here are concrete scenarios and tips tailored to common study designs.

Scenario 1: Discovery in a well-characterized cell line (HMW DNA)

Goal: Broad profiling of eccDNA repertoire to generate hypotheses.
Recommended: Exonuclease digestion → RCA → short-read sequencing; long-read follow-up for circles >10 kb.
Tips: Run three technical replicates to assess reproducibility. Use spike-in circles at 0.1×, 1×, 10× ratios to calibrate bias. Keep RCA to 3 hours to limit coverage spikes.
Analysis: Use Circle-Map with split-read ≥2 threshold and circle score filtering; report length distribution and genomic annotations.

Scenario 2: Discovery in low-input biopsy/FFPE

Goal: Detect eccDNA in scarce, fragmented samples.
Recommended: RCA-first approach with careful time limitation; optional mild exonuclease cleanup if integrity allows.
Tips: Opt for low-input library kits; accept increased small-circle bias; plan targeted validation for key loci via amplicon/Junction PCR.
Analysis: Compare with non-RCA control if possible; quantify background fraction and duplication rates.

Scenario 3: Targeted validation of suspected oncogene eccDNA

Goal: Confirm and quantify circles at specific loci (e.g., MYC).
Recommended: Exonuclease-cleaned library → hybrid capture panel including junctions and flanks; avoid RCA to minimize bias.
Tips: Design probes with redundancy around breakpoints; validate with junction PCR/Sanger; if complex, perform nanopore targeted long-read.
Analysis: Focus on on-target fraction, coverage uniformity, and junction-spanning reads; report copy-number estimates with caution.

Scenario 4: Rapid quantification for a deadline (conference/poster)

Goal: Produce a clear, reproducible signal at one or two loci.
Recommended: Amplicon sequencing and junction PCR/Sanger on exonuclease-cleaned DNA; optional long-read for structural clarity.
Tips: Keep workflow simple; prioritize controls and clear reporting; avoid RCA unless input forces it.
Analysis: Provide direct junction confirmation and replicate concordance; include spike-in recovery metrics in the poster.

Practical tips across scenarios

Pilot first: Titrate enzyme units, RCA times, and bead ratios with a small batch; lock parameters before scaling.
Document rigorously: Record lot numbers, incubation times, QC checkpoints; reproducibility is your publication insurance.
Plan orthogonal validation: Combine junction PCR/Sanger with long-read confirmation for key findings.
Budget for depth: Discovery studies benefit from 50–100M PE reads; targeted panels often suffice at 5–20M PE reads.

Conclusion: Choose by Discovery vs Targeted

If your objective is discovery, start with exonuclease digestion followed by RCA, accept small-circle bias as the cost of sensitivity, and plan orthogonal validation (junction PCR, long-read confirmation) for key findings. If your goal is targeted validation or quantification, prioritize hybrid capture on exonuclease-cleaned libraries, incorporate junction-specific assays, and keep your QC tight and reproducible.

If you need a neutral implementation partner for targeted capture or confirmatory sequencing, CD Genomics supports RUO projects with custom panels and long-read confirmation. Genomics Services.

Author

Yang H. — Senior Scientist, CD Genomics; University of Florida.

Yang is a genomics researcher with over 10 years of research experience in genetics, molecular and cellular biology, sequencing workflows, and bioinformatic analysis. Skilled in both laboratory techniques and data interpretation, Yang supports RUO study design and NGS-based projects.

References:

Comparative analysis of eccDNA methodologies and pipeline thresholds (2024): Gao et al., PMC11502736.
Genome-wide characterization of eccDNA with RCA-derived workflows (2023): Jiang et al., PMC10300174.
ATAC-derived detection of eccDNA and validation (2021): Su et al., PMC8161110.
NEB Exonuclease V (RecBCD) reagent page: NEB M0345.
NEB phi29-XT RCA kit and isothermal amplification guidance: NEB E1603.

Soft CTA: Contact CD Genomics to discuss RUO targeted panels or long-read confirmation for eccDNA studies: Genomics Services.

For research purposes only, not intended for clinical diagnosis, treatment, or individual health assessments.