AAV Integration Quantification vs Integration Site Mapping: Which Analysis Does Your Study Need?

Figure 1. Quantification and site mapping answer different AAV integration questions and should be selected by the intended claim.

AAV integration quantification asks how much integration-associated signal is present in a defined sample, whereas integration site mapping asks where vector–host junctions occur and how those sites are distributed. A study may need one endpoint, both endpoints, or a staged design in which site discovery is followed by locus-specific measurement. This guide defines the questions each analysis can support, the evidence it cannot supply on its own, and the sample, control, sequencing, and validation choices that make results interpretable.

This discussion concerns research-use genomic analysis. CD Genomics does not offer these services for clinical diagnosis or individual treatment decisions.

Key Takeaways

• Choose the endpoint before the assay. A percentage or copy estimate does not identify genomic coordinates, and a list of coordinates does not automatically measure total integrated burden.
• Define the denominator. Quantification may be reported per input genome, cell equivalent, vector copy, or analyzed molecule; those values are not interchangeable.
• Treat recovered sites as a sampled distribution. Junction recovery depends on DNA input, fragmentation, enrichment, amplification, mapping, and duplicate handling.
• Use controls to distinguish integration from background. Residual vector DNA, chimeric library molecules, index carryover, and ambiguous host mapping can all create misleading signals.
• Combine methods when the biological claim spans amount and location. Orthogonal assays can resolve disagreements between a global estimate and junction-level evidence.

Two Questions That Are Often Confused

The phrase "AAV integration analysis" can describe several distinct endpoints. Some projects need a population-level estimate of integration-associated vector sequence. Others need genomic coordinates, nearby genes, orientation, clonality patterns, or junction sequence. A third group needs structural reconstruction of integrated vector fragments. No single output should be assumed to answer all of these questions.

The AAV Integration Site Analysis service is oriented toward vector–host junction discovery and genomic annotation. By contrast, a quantitative endpoint must be tied to a defined molecular denominator and a validated measurement model. Projects that begin without this distinction often produce technically correct data that cannot support the intended conclusion.

What AAV Integration Quantification Measures

Quantification reduces a complex molecular population to a numerical estimate. The estimate may represent integration-positive genomes per host genome equivalent, vector copies associated with high-molecular-weight DNA, or the abundance of a validated junction. Before interpreting the number, the report must specify exactly what molecular species the assay detects.

Global and locus-specific quantification are different

A global estimate aims to measure integration-associated vector material across all loci. Depending on the workflow, it may use enrichment of high-molecular-weight DNA, depletion of episomal forms, calibrated PCR, sequencing, or a combined model. A locus-specific assay instead measures one known vector–host junction. It is useful for tracking a previously discovered event but cannot reveal unrelated integration sites.

Vector-targeted qPCR or digital PCR alone generally measures the targeted vector sequence, not its genomic state. Episomal monomers, concatemers, partial vector fragments, and integrated copies may share that target. A method intended to quantify integration must explain how nonintegrated vector DNA is excluded, modeled, or independently measured. The AAV genome integrity metrics guide is useful when the parallel question concerns packaged vector genomes rather than host integration.

The denominator controls meaning

"Integration frequency" is incomplete without a denominator. Per-cell, per-diploid-genome, per-analyzed-genome, per-vector-genome, and per-microgram-DNA estimates describe different quantities. Tissue ploidy, cell mixture, extraction recovery, and DNA damage can further separate the experimental denominator from the biological one.

Useful quantitative reporting therefore includes:

• the vector sequence measured and why it represents the intended molecular species;
• the host reference used to estimate genome equivalents;
• the calibration material, dynamic range, limit of blank, and limit of quantification;
• the treatment of values below quantifiable range;
• replicate-level results rather than only a pooled mean;
• an uncertainty interval and the assumptions used to calculate it.

Quantification is strongest when the assay is challenged with negative matrix, vector-spiked material, integrated positive controls, and samples containing abundant nonintegrated vector DNA. These controls reveal whether the workflow measures integration or merely vector presence.

What Integration Site Mapping Measures

Site mapping searches for molecules that contain both vector and host sequence. A high-confidence site should have a defined host coordinate, vector boundary, strand, supporting molecule count, and mapping-quality context. Ideally, the evidence includes independently observed junction molecules rather than many amplification copies of the same starting molecule.

Target-enrichment, ligation-mediated, sonication-based, and long-read workflows can recover different subsets of sites. Comparative studies show that method choice changes sensitivity, specificity, and the types of structures observed [1,2]. The comparison of short-read and long-read AAV sequencing provides additional platform context, but read length is only one component of site recovery.

A site list is not a census

Every mapping workflow has a sampling function. Restriction-site location, fragment length, bait placement, PCR efficiency, vector-boundary variability, host mappability, and sequencing depth all influence which junctions are recovered. Consequently, the number of unique mapped sites should not be interpreted as the number of integration-bearing cells without a validated statistical and molecular model.

Molecular barcodes or random shearing coordinates can help identify independent molecules, but deduplication rules must be stated. Two reads with the same coordinate may be PCR duplicates, independent molecules from an expanded clone, or repeated sampling of a prevalent event. Those possibilities cannot be separated from read counts alone.

Annotation is downstream of evidence

After site calling, coordinates can be annotated against genes, regulatory regions, repeat classes, chromatin features, or other genomic tracks. Enrichment claims require an appropriate background model that reflects the assay's recoverable genome, not merely the full genome. Reference build, annotation release, interval definitions, and multiple-testing correction should be recorded.

Figure 2. Integration site mapping converts junction-bearing molecules into filtered coordinates and annotated site distributions.

Quantification Versus Mapping at a Glance

The following comparison helps identify which result belongs in the primary endpoint and which belongs in supportive analysis.

Researchers evaluating a broader vector program can use Viral Vector Development Solutions to place integration endpoints alongside vector identity and quality questions. For sequence-level assessment of the vector preparation itself, AAV Sequencing addresses a different sample and evidence layer.

A Decision Workflow for Study Planning

Start with the claim the study must support. If the intended conclusion contains "how much," quantification is primary. If it contains "where," "which genes," or "what distribution," mapping is primary. If it contains both, plan two compatible evidence streams rather than expecting one pipeline to provide a complete answer.

1. Write the endpoint as a sentence. Include the sample, comparison, unit, time point, and acceptable uncertainty.
2. Choose the molecular population. Decide whether the analysis concerns total tissue DNA, a sorted population, a clonal culture, or another defined material.
3. Estimate expected signal. Low-frequency events need adequate genome equivalents and independently processed replicates, not only deeper sequencing of a small library.
4. Select controls before samples are consumed. Include biological negatives, process blanks, and positive controls that challenge the key failure mode.
5. Predefine calling and confirmation rules. State read support, mapping quality, replicate concordance, and orthogonal validation thresholds.
6. Match deliverables to decisions. A site table, a quantitative estimate, a structural map, and a validation record should not be merged into an undefined "integration result."

When quantification alone may be sufficient

A quantitative-only design may fit an exploratory comparison in which the location of individual sites is not part of the hypothesis. It can also support tracking of a previously validated junction across research samples. The assay must still demonstrate specificity for the molecular form being measured.

When mapping alone may be sufficient

Mapping may be the primary analysis when the objective is to identify genomic contexts, compare site distributions, or nominate loci for validation. It is not sufficient if the conclusion requires a total integration rate. A mapping workflow can report recovery depth and molecule counts while explicitly avoiding conversion to a per-cell burden.

When both are needed

Both analyses are appropriate when a study must relate global burden to location, compare tissues that differ in total signal and site diversity, or follow selected integration sites over time. In such designs, quantification establishes scale, mapping establishes distribution, and targeted validation confirms selected junctions.

Samples, Controls, and Replicates

Integration analysis is often limited by informative starting molecules. Sequencing the same amplified library more deeply cannot replace missing biological input. The sample plan should therefore be built around genome equivalents, expected event frequency, tissue heterogeneity, and DNA integrity.

For tissue studies, separate biological replicates preserve between-subject variation. Pooling may increase input but removes the ability to estimate variability and can make a prevalent site appear more general than it is. For longitudinal designs, consistent sampling, extraction, and input normalization are essential because technical shifts can resemble biological changes.

The guide to validating low-frequency AAV integration sites explains why read support, independent libraries, and orthogonal assays should be combined for rare events. A low-frequency claim should include the number of genomes analyzed and the probability of observing the event, not only the nominal sequencing depth.

Figure 3. Controls and replicate structure should target the distinct failure modes of quantitative and site-mapping workflows.

Selecting Sequencing and Validation Methods

Short-read junction capture offers efficient, high-throughput site discovery when vector boundaries are known. Long-read enrichment can connect longer vector and host segments, distinguish complex structures, and reduce dependence on reconstruction from short fragments [2,3,6]. Whole-genome sequencing supplies broader host-genome context but may have limited sensitivity for rare integration-bearing molecules at practical depth.

No platform eliminates the need for orthogonal confirmation. Junction PCR and Sanger sequencing can verify a nominated boundary. Digital PCR can quantify a validated junction or vector target. Long-read sequencing can test whether distant components reside on the same molecule. The AAV confirmation guide for PCR, Sanger sequencing, and NGS helps align confirmation method with claim.

Recent studies have combined or cross-validated methods because integration frequency and structure can be difficult to infer from one data type [1,2,4–6]. This is especially important when the result depends on rare junctions, complex concatemers, or a large excess of nonintegrated vector genomes.

What the Final Report Should Contain

A decision-ready report separates measured values from inferred interpretations. It also preserves enough method information for another analyst to understand how sites and quantities were generated.

• For quantification: assay target, molecular-state strategy, denominator, genome-equivalent calculation, calibration model, replicate values, quantifiable range, uncertainty, and treatment of censored results.
• For mapping: host and vector references, junction coordinates, strand, vector boundary, molecule-aware support, mapping quality, blacklist or repeat context, replicate occurrence, and confirmation status.
• For distribution analysis: annotation source, recoverable-genome background, statistical test, multiple-testing method, and minimum site count.
• For combined studies: a clear rule linking global measurements to site-level observations without treating one as a direct substitute for the other.

The AAV integration analysis overview provides broader context for method selection. If integration-site distribution is the principal deliverable, the report should retain site-level evidence. If total burden is principal, raw site counts should remain supportive rather than becoming the denominator.

Figure 4. A combined report keeps quantitative scale, genomic locations, distribution statistics, and validation status as distinct evidence layers.

FAQ

For research use only. Not for use in diagnostic procedures or individual treatment decisions.