Chimeric Reads in Long-Read Sequencing: Causes, Detection, and Impact on Alignment Accuracy

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-21

Key Takeaways

• Chimeric reads, formed by joining distinct genomic segments within a single long read, are significant artifacts in PacBio and Oxford Nanopore sequencing, arising from library preparation (ligation artifacts, concatemer formation) and amplification (MDA mis-priming).
• Detection relies on identifying supplementary alignments in SAM/BAM files (SAM flag 2048) and employing specialized tools like Breakinator for foldback artifacts and 3rd-ChimeraMiner for restoring MDA-induced chimeras.
• The impact of chimerism is substantial, leading to false-positive structural variant calls (e.g., inversions, translocations) and misassembled contigs, with one study showing restoration of MDA chimeras removed ~97% of inversion calls.
• Practical mitigation involves optimizing adapter ligation, careful size selection, limiting MDA amplification time and rounds, and using foldback-aware filtering before downstream analyses like variant calling or genome assembly.
• Barcode-associated recombination in mutagenized libraries requires specific tools like Pacybara to cluster reads by barcode similarity and detect recombinant clones, preventing false genotype-barcode associations.

Chimeric reads are single sequencing molecules that contain sequence segments originating from two or more distinct genomic locations joined together in one read. In long-read sequencing platforms such as PacBio and Oxford Nanopore, chimeric reads arise from library preparation artifacts, amplification errors, and occasionally from genuine structural rearrangements in the sample genome. These artifacts pose a significant problem for downstream analysis because aligners may map different segments of a chimeric read to different genomic positions, producing supplementary alignments that mimic structural variants. For researchers analyzing structural variation, copy number changes, or assembling genomes from long reads, failing to identify and handle chimeric reads can lead to false-positive variant calls, misassembled contigs, and incorrect biological conclusions. This article describes the known sources of chimeric reads in PacBio and Nanopore libraries, explains how to detect them using alignment flags and specialized tools, and provides practical strategies to minimize their impact on alignment accuracy and downstream interpretation.

At a Glance

The table below summarizes the primary causes of chimeric reads, their detection methods, and the practical actions researchers can take to reduce their influence on long-read analysis.

Understanding Chimeric Reads in Long-Read Sequencing

Definition and Molecular Basis

A chimeric read is a sequencing molecule whose sequence does not correspond to a contiguous region of the source genome. Instead, the read contains segments that originate from different genomic loci that have been joined during sample preparation or sequencing. The junction points between these segments are not present in the original genome and therefore represent technical artifacts instead of biological variation.

In short-read sequencing, chimeric reads are relatively rare and often discarded during quality filtering because the read lengths are too short to span multiple genomic regions with high confidence. Long-read sequencing changes this dynamic. PacBio and Oxford Nanopore reads routinely exceed 10 kilobases, and some platforms can produce reads over 20 kilobases. A chimeric read of this length can contain two or more complete segments from different genomic locations, each long enough to align uniquely. When an aligner processes such a read, it may split the alignment into multiple segments, each mapping to a different locus. These split alignments appear in the SAM/BAM output as primary and supplementary alignments, and they can be misinterpreted as evidence of structural variation.

The practical consequence is that chimeric reads introduce false structural variant calls, particularly inversions, deletions, and translocations. A study of MDA-amplified samples sequenced on the PacBio platform found that the vast majority of recognized chimeric sequences were artifacts instead of genuine genomic structures, and restoring these chimeras to their original structures removed approximately 97% of inversion calls on average (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). This finding underscores the need for systematic chimera detection in long-read analysis workflows.

Why Long Reads Are Especially Vulnerable

Long reads offer substantial advantages for genome assembly, structural variant detection, and phasing, but these same advantages create vulnerability to chimerism. The probability that a read contains a chimeric junction increases with read length because longer molecules have more opportunities to incorporate sequence from multiple loci. Additionally, the library preparation steps required to generate long reads often involve enzymatic reactions, size selection, and amplification that can create artificial junctions.

The error profile of long-read platforms also complicates chimera detection. PacBio and Oxford Nanopore reads have higher per-base error rates than Illumina short reads, although PacBio HiFi reads achieve higher accuracy through circular consensus sequencing. When a chimeric junction occurs in a region with sequencing errors, distinguishing the artifact from a genuine structural variant becomes more difficult. Alignment-based detection tools must therefore account for the expected error rates of the platform and chemistry being used.

Sources of Chimeric Reads in PacBio and Nanopore Libraries

Ligation Artifacts and Adapter Concatenation

Library preparation for long-read sequencing typically involves attaching adapters to DNA fragments. When adapter ligation is inefficient or when fragment ends are damaged, multiple DNA molecules can ligate together in a single molecule. This concatemer formation produces a sequencing template that contains sequence from two or more original fragments joined end to end. During sequencing, the entire concatemer may be read as a single molecule, producing a chimeric read.

Ligation artifacts are more common when input DNA is fragmented or when adapter-to-insert ratios are suboptimal. Size selection can reduce but not eliminate concatemers, because molecules of similar length can still contain internal junctions. Researchers who observe a high proportion of chimeric reads in their libraries should examine their ligation conditions, adapter concentrations, and size-selection parameters as potential contributors.

Multiple Displacement Amplification Mis-Priming

Multiple displacement amplification (MDA) using phi29 DNA polymerase has become a common method for whole genome amplification, particularly when sample quantity is limited. MDA generates high molecular weight DNA with broad genome coverage, and coupling MDA with long-read sequencing enables sequencing of amplicons exceeding 20 kilobases. However, MDA is prone to forming chimeric sequences through mis-priming events during amplification.

A detailed analysis of MDA amplicons sequenced on the PacBio platform revealed that mis-priming events occur more frequently than widely recognized. The proportion of chimeric sequences accumulated from 42% to over 78% as amplification continued, meaning that longer amplification times produced progressively more chimeric molecules. Critically, 99.92% of recognized chimeric sequences in that study were demonstrated to be artifacts formed during MDA instead of structures present in the original genomes (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). This finding has direct implications for researchers who amplify limited samples before long-read sequencing: the amplification process itself generates chimeras that will appear in the sequencing data and interfere with structural variant analysis.

The practical takeaway is that MDA-amplified samples require dedicated chimera detection and restoration steps before structural variant calling. Tools such as 3rd-ChimeraMiner can recognize chimeric structures and restore them to their original configurations, recycling supplementary alignments that would otherwise introduce false-positive structural variants (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Foldback Artifacts

Foldback artifacts represent a distinct class of chimeric structures in which a read contains sequence that folds back on itself, creating an inverted repeat or hairpin-like structure. These artifacts have been observed in both Oxford Nanopore and PacBio data across a range of specimens, library types, sequencing chemistries, sequencing machines, and base-calling software (Detecting Foldback Artifacts in Long-reads).

Foldback artifacts are particularly problematic because they can be mistaken for genuine inverted repeats or complex structural variants. Standard quality control tools may not flag these reads because their overall quality metrics appear normal. Alignment-based detection approaches, such as those implemented in the Breakinator tool, can identify foldback artifacts by examining the alignment patterns for evidence of self-complementary or inverted sequence segments (Detecting Foldback Artifacts in Long-reads).

The occurrence of foldback artifacts varies across library types and sequencing platforms, so researchers should profile their specific workflow to understand the baseline rate. Comparing results across different library preparation kits, sequencing chemistries, and base-calling software versions can reveal whether foldback artifacts are introduced at a particular step (Detecting foldback artifacts in long-reads).

Barcode-Associated Recombination in Mutagenized Libraries

For applications that use barcoded libraries, such as multiplexed assays of variant effects (MAVEs), chimeric reads present an additional challenge. In these workflows, long reads are used to associate barcodes with genotypes, and chimeric molecules can create false associations between a barcode and a genotype that does not actually correspond to that barcode (Pacybara: Accurate long-read sequencing for barcoded mutagenized allelic libraries).

The Pacybara tool was developed specifically to address this problem in barcoded mutagenized allelic libraries. Pacybara clusters long reads based on the similarity of error-prone barcodes, detects barcodes that have been associated with multiple genotypes, and identifies recombinant (chimeric) clones. The tool also reduces false positive indel calls that can arise from sequencing errors or chimeric molecules (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries). For researchers using long-read sequencing of barcoded libraries, incorporating chimera detection at the barcode-genotype association step is essential for accurate variant effect mapping.

Detection of Chimeric Reads Using Alignment Flags

Primary and Supplementary Alignments in SAM/BAM Format

The SAM (Sequence Alignment/Map) format provides the standard representation of read alignments. When a read aligns to multiple genomic locations, the aligner designates one alignment as primary and others as supplementary. Supplementary alignments are recorded with the SAM flag 2048 (0x800), and they indicate that the read has been split into multiple segments that map to different positions.

For chimeric reads, the presence of supplementary alignments is a key indicator. A read that produces two or more alignments to distant genomic locations is likely chimeric, although genuine structural variants can also produce split alignments. The distinction between artifact and biology requires additional evidence, such as the orientation of the aligned segments, the distance between them, and whether the junction is supported by multiple independent reads.

Researchers should routinely inspect the supplementary alignment flags in their BAM files as a first-pass chimera screen. Tools that parse SAM/BAM files, including those available through Bioconductor packages, can extract reads with supplementary alignments and examine their mapping patterns. The Bioconductor project provides official documentation for packages that handle alignment data and can be used to build chimera detection workflows.

Alignment-Based Detection Tools

Alignment-based detection tools examine the pattern of alignments across a read to identify chimeric structures. The Breakinator tool uses an alignment-based approach to flag putative foldback artifact reads and previously known chimeric artifacts. Importantly, Breakinator can detect artifacts that are missed by existing quality control tools, making it a valuable addition to standard QC pipelines (Detecting Foldback Artifacts in Long-reads).

The development of Breakinator involved profiling the occurrence of foldbacks and chimeric reads in both Oxford Nanopore and PacBio sequences across a range of specimens, library types, sequencing chemistries, sequencing machines, and base-calling software (Detecting foldback artifacts in long-reads). This profiling provides a reference for researchers who want to understand the expected rates of these artifacts in their own data.

For MDA-amplified samples, the 3rd-ChimeraMiner pipeline offers a different approach: it recognizes chimeric sequences and restores them to their original structures. This restoration approach is particularly useful for structural variant analysis because it recycles supplementary alignments that would otherwise contribute false-positive calls. The pipeline was constructed specifically for long-read sequencing data and has been applied to five long-read datasets and one high-fidelity long-read dataset with various amplification folds (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Distinguishing Chimeric Artifacts from Genuine Structural Variants

The central challenge in chimera detection is distinguishing technical artifacts from genuine structural variants. Both produce split alignments, and both can involve distant genomic regions. Several lines of evidence can help make this distinction:

Read-level evidence. A chimeric artifact is often present in a single read or a small number of reads, whereas a genuine structural variant should be supported by multiple independent reads. Examining the read depth and the number of supporting reads at a putative junction can help differentiate artifacts from biology.

Junction characteristics. Chimeric artifacts from ligation or amplification often have junction points that do not correspond to known genomic features. Genuine structural variants may have junction points at repetitive elements, segmental duplications, or other genomic features that promote rearrangement.

Orientation patterns. Foldback artifacts produce characteristic orientation patterns in which the sequence folds back on itself. These patterns can be recognized by examining the strand and orientation of the aligned segments (Detecting Foldback Artifacts in Long-reads).

Consistency across samples. If the same putative variant appears in multiple samples from the same library preparation batch, it may be a batch-specific artifact. Conversely, a variant that appears consistently across independent libraries is more likely to be genuine.

Researchers should apply multiple lines of evidence before classifying a split alignment as a genuine structural variant. The cost of a false-positive call is wasted validation effort, while the cost of a false-negative call is a missed biological finding.

Practical Workflow for Chimera Detection and Management

Step 1: Assess Library Preparation and Amplification History

Before analyzing sequencing data, document the library preparation method and any amplification steps. This information determines the expected types of chimeric artifacts. Key questions include:

• Was the sample amplified before sequencing? If so, what method was used and for how many cycles or hours?
• Was MDA used? If so, the expected chimera proportion may be substantial, particularly if amplification was prolonged (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).
• What library preparation kit and protocol were used? Different kits have different propensities for ligation artifacts.
• What size selection method was applied? Size selection can reduce but not eliminate concatemers.

This documentation should be recorded in the project metadata and considered when interpreting structural variant calls.

Step 2: Run Standard Quality Control and Inspect Alignment Flags

After alignment, run standard quality control metrics and inspect the distribution of alignment flags. Calculate the proportion of reads with supplementary alignments and examine whether these reads cluster in particular genomic regions or library preparation batches.

The Galaxy Training Network provides accessible tutorials for quality control and alignment workflows that can be adapted for long-read data. These tutorials cover the practical steps of inspecting alignment files and identifying problematic reads.

Step 3: Apply Chimera-Specific Detection Tools

Based on the library preparation history, apply appropriate chimera detection tools:

• For MDA-amplified samples, run 3rd-ChimeraMiner to recognize and restore chimeric structures (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).
• For general chimera and foldback detection, run Breakinator to flag putative artifacts (Detecting Foldback Artifacts in Long-reads).
• For barcoded mutagenized libraries, run Pacybara to detect recombinant clones and correct barcode-genotype associations (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries).

These tools should be integrated into the analysis pipeline before structural variant calling or genome assembly. Running them after variant calling may allow false-positive calls to propagate into downstream analyses.

Step 4: Compare Results Across Conditions

If chimeric reads are detected, investigate whether they correlate with specific library preparation batches, sequencing runs, or base-calling software versions. The profiling performed during Breakinator development showed that artifact occurrence varies across these conditions, so understanding the specific conditions of your data can help identify the source of the problem (Detecting foldback artifacts in long-reads).

For example, if chimeric reads are concentrated in samples processed on a particular sequencing machine or with a particular chemistry, the issue may be instrument-specific. If they correlate with a particular library preparation kit, the issue may be protocol-specific.

Step 5: Document and Report Chimera Rates

Record the proportion of chimeric reads in each sample and include this information in methods sections and data repositories. This documentation allows other researchers to assess the quality of the data and to compare results across studies. It also provides a baseline for monitoring library preparation quality over time.

The NCBI Data Resources provide official repositories for sequence data and associated metadata. Including chimera detection metrics in the metadata for deposited datasets improves the utility of those datasets for secondary analysis.

Records and Measurements for Chimera Monitoring

Key Metrics to Track

Establish a set of metrics to monitor chimera rates across samples and experiments. These metrics should be recorded consistently and reviewed regularly:

Proportion of reads with supplementary alignments. This is the most accessible metric and can be calculated from the BAM file. A sudden increase in this proportion may indicate a library preparation problem.

Chimera proportion by amplification round. For MDA-amplified samples, track the chimera proportion as a function of amplification time or cycle number. The finding that chimera proportion accumulates from 42% to over 78% with continued amplification provides a reference range, although the exact values will depend on the specific protocol (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Foldback artifact rate. If using Breakinator, record the number of reads flagged as foldback artifacts. This rate should be compared across library types and sequencing conditions (Detecting Foldback Artifacts in Long-reads).

Barcode-genotype discordance rate. For barcoded libraries, track the proportion of barcodes associated with multiple genotypes, as this may indicate chimeric molecules or barcode errors (Pacybara: Accurate long-read sequencing for barcoded mutagenized allelic libraries).

False-positive structural variant rate. After applying chimera detection and restoration, compare the number of structural variant calls before and after filtering. The finding that restoring chimeras removed 97% of inversions on average in MDA-amplified samples illustrates the magnitude of this effect (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Record-Keeping Practices

Maintain a laboratory notebook or electronic record that documents:

• Library preparation protocol details, including any amplification steps
• Sequencing platform, chemistry, and base-calling software versions
• Alignment software and parameters
• Chimera detection tools and versions
• The number of reads flagged as chimeric or foldback artifacts
• The number of structural variant calls before and after chimera filtering

These records enable troubleshooting when problems arise and provide the information needed for methods sections in publications.

Common Failure Patterns in Chimera Management

Failure to Detect Chimeras in Amplified Samples

A common failure is assuming that standard quality control metrics will identify chimeric reads. Standard QC tools may not flag chimeric reads because their quality scores and length distributions appear normal. The finding that Breakinator can detect artifacts missed by existing quality control tools highlights this limitation (Detecting Foldback Artifacts in Long-reads). Researchers who skip chimera-specific detection may unknowingly include chimeric reads in their variant calls.

Misinterpreting Supplementary Alignments as Structural Variants

Another failure pattern is treating all supplementary alignments as evidence of structural variation. In MDA-amplified samples, the vast majority of recognized chimeric sequences are artifacts, and restoring them to their original structures removes most inversion calls (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). Researchers who do not perform this restoration will report false-positive structural variants that cannot be validated.

Overlooking Foldback Artifacts

Foldback artifacts are less well known than other chimeric structures and may be missed by both standard QC and general chimera detection tools. The development of Breakinator specifically to flag foldback artifacts indicates that these structures were causing problems in real analyses (Detecting Foldback Artifacts in Long-reads). Researchers should specifically check for foldback artifacts, particularly when analyzing data for complex structural variants.

Ignoring Barcode-Genotype Discordance

In barcoded library experiments, chimeric molecules can create false barcode-genotype associations. If these associations are not detected and corrected, the resulting genotype-phenotype maps will contain errors. Pacybara was developed to address this problem, and its use should be considered standard practice for barcoded mutagenized library analysis (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries).

Applying a One-Size-Fits-All Filtering Approach

Different types of chimeric artifacts require different detection and management strategies. A filtering approach that works for ligation artifacts may not address foldback artifacts or MDA-induced chimeras. Researchers should tailor their chimera management to the specific library preparation and amplification methods used.

Limitations of Chimera Detection and Management

Detection Tools Have Known Boundaries

Chimera detection tools operate within defined boundaries. Alignment-based tools require that the chimeric segments be long enough and unique enough to align to distinct genomic locations. If a chimeric segment is too short or falls in a repetitive region, the aligner may not produce a supplementary alignment, and the chimera will go undetected.

Similarly, restoration tools such as 3rd-ChimeraMiner can restore chimeric structures to their original configurations, but the accuracy of restoration depends on the quality of the alignments and the complexity of the chimeric structure. Highly complex chimeras involving multiple segments may not be fully resolved (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Amplification-Induced Chimeras Cannot Be Fully Eliminated

For samples that require amplification, chimeric molecules are an inherent byproduct of the amplification process. The finding that chimera proportion accumulates with continued amplification means that researchers face a tradeoff between the amount of DNA produced and the proportion of chimeric molecules (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). Reducing amplification time can reduce chimeras but may not produce sufficient DNA for sequencing.

Platform and Chemistry Differences Affect Generalizability

The rates of chimeric and foldback artifacts vary across sequencing platforms, chemistries, and base-calling software (Detecting foldback artifacts in long-reads). Findings from one platform or chemistry may not directly transfer to another. Researchers should profile their specific workflow to establish baseline rates and should be cautious about comparing chimera rates across studies that used different platforms.

Structural Variant Validation Remains Necessary

Even with careful chimera detection and management, structural variant calls from long-read data should be validated using orthogonal methods. Chimeric artifacts can produce patterns that mimic genuine variants, and the distinction is not always clear. Validation approaches may include PCR amplification across the putative junction, optical mapping, or sequencing with a different platform.

Professional Escalation Criteria

When to Seek Additional Expertise

Researchers should consider consulting with bioinformatics specialists or sequencing facility staff when:

• The proportion of reads with supplementary alignments exceeds the expected range for the library preparation method
• Chimera detection tools flag an unexpectedly high number of reads
• Structural variant calls cannot be validated despite multiple attempts
• The same chimeric pattern appears across multiple independent libraries, suggesting a systematic protocol issue
• Barcode-genotype discordance rates are high in mutagenized library experiments

When to Revisit Library Preparation

If chimera rates are consistently high across multiple samples, the library preparation protocol should be revisited. Potential adjustments include:

• Optimizing adapter ligation conditions to reduce concatemer formation
• Reducing amplification time or rounds for MDA-amplified samples
• Changing size selection parameters to exclude molecules with internal junctions
• Testing alternative library preparation kits

When to Consider Alternative Sequencing Approaches

For samples that consistently produce high chimera rates, alternative sequencing approaches may be warranted. Options include:

• Sequencing unamplified material when sample quantity permits
• Using high-fidelity long-read sequencing, which may have different artifact profiles
• Combining long-read sequencing with short-read sequencing for cross-validation of structural variant calls

Decision Framework for Selecting Chimera Management Strategies

Matching the Management Approach to the Analysis Goal

The choice between removing chimeric reads, restoring them to original structures, or applying read-level filtering depends on the downstream analysis objective. A single management strategy does not fit all long-read applications, and applying the wrong approach can either discard useful data or allow false-positive variants to persist. This section provides a practical decision framework that connects the analysis goal, the library preparation history, and the available detection tools to a specific management action.

Decision Point 1: Define the Primary Downstream Analysis

Before selecting a chimera management strategy, document the primary analysis that will be performed on the aligned reads. The three most common long-read applications have different tolerance for chimeric reads and different requirements for data preservation.

Genome assembly. Chimeric reads cause misassemblies because assemblers may use the chimeric junction as evidence of adjacency between genomic regions that are not actually adjacent. For assembly, the priority is removing chimeric reads or splitting them at the chimeric junction before assembly. Restoration to original structures is less useful because the assembler needs contiguous, non-chimeric sequences. Reads flagged as chimeric should be excluded from the assembly input or trimmed at the detected junction points.

Structural variant calling. Chimeric reads produce false-positive structural variant calls, particularly inversions and translocations. For this application, restoration is often preferable to removal because restoration recycles the non-chimeric portions of the read and preserves coverage at the true genomic loci. The 3rd-ChimeraMiner pipeline was designed for this purpose, restoring chimeric sequences to their original structures and removing the vast majority of supplementary alignments that introduce false-positive structural variants (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). In MDA-amplified samples, this restoration approach removed 97% of inversion calls on average.

Barcode-genotype association in mutagenized libraries. For multiplexed assays of variant effects and similar barcoded library applications, the critical requirement is accurate association between barcodes and genotypes. Chimeric molecules create false associations that corrupt the genotype-phenotype map. The management strategy here is not simply removing or restoring reads but detecting and correcting the barcode-genotype associations themselves. Pacybara addresses this by clustering reads based on barcode similarity, detecting barcodes associated with multiple genotypes, and identifying recombinant clones (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries).

Decision Point 2: Assess the Library Preparation and Amplification History

The expected type and proportion of chimeric reads depend on the library preparation method. Document the following before selecting a management strategy:

Was multiple displacement amplification used? If yes, the chimera proportion may be substantial and increases with amplification time. One study found that the proportion of chimeric sequences accumulated from 42% to over 78% as amplification continued (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). For MDA-amplified samples, restoration-based approaches such as 3rd-ChimeraMiner are strongly preferred because removal would discard a large fraction of the data.

Was the sample amplified at all? Unamplified samples have lower chimera rates, and simple removal of flagged reads may be sufficient. The tradeoff between data loss and artifact removal is less severe when the baseline chimera rate is low.

What library preparation kit and protocol were used? Different kits have different propensities for ligation artifacts and concatemer formation. If the kit is known to produce concatemers, the management strategy should include inspection of supplementary alignment flags as a routine step.

What sequencing platform and chemistry were used? Foldback artifacts have been observed in both Oxford Nanopore and PacBio data across a range of specimens, library types, sequencing chemistries, sequencing machines, and base-calling software (Detecting Foldback Artifacts in Long-reads). The specific platform and chemistry affect the expected artifact profile and should inform the choice of detection tools.

Decision Point 3: Select the Detection Tool Based on Artifact Type

The decision framework should map the expected artifact type to the appropriate detection tool:

Ligation artifacts and concatemers. These produce reads with segments from different genomic locations joined end to end. Detection relies on inspecting SAM/BAM supplementary alignment flags. Reads with multiple alignments to distant loci are candidates. Standard alignment inspection workflows available through Bioconductor packages can extract these reads for further examination.

MDA-induced chimeras from mis-priming. These require dedicated detection and restoration. The 3rd-ChimeraMiner pipeline recognizes chimeric structures and restores them to their original configurations (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). This tool should be applied specifically to MDA-amplified samples before structural variant calling.

Foldback artifacts. These create hairpin-like structures with inverted sequence patterns. Standard quality control tools may miss them, and alignment-based detection is required. Breakinator was developed specifically to flag putative foldback artifact reads and previously known chimeric artifacts, and it can detect artifacts missed by existing quality control tools (Detecting Foldback Artifacts in Long-reads).

Barcode-associated recombination. For barcoded libraries, the detection and correction must occur at the barcode-genotype association step. Pacybara handles this by clustering reads based on barcode similarity and detecting barcodes associated with multiple genotypes (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries).

Decision Point 4: Determine the Management Action

Based on the analysis goal and the detected artifact type, select one of the following management actions:

Action A: Remove flagged reads. This is appropriate for genome assembly and for analyses where the proportion of chimeric reads is low. Removal prevents chimeric junctions from influencing assembly graphs or variant calls. The cost is loss of coverage at the true loci, which may be acceptable when chimera rates are below 5%.

Action B: Restore chimeric reads to original structures. This is appropriate for structural variant calling in amplified samples where chimera rates are high. Restoration preserves coverage and recycles supplementary alignments that would otherwise contribute false-positive calls. The 3rd-ChimeraMiner pipeline implements this approach for long-read data (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Action C: Correct barcode-genotype associations. This is appropriate for barcoded mutagenized libraries. The management action is not read removal but correction of the associations that chimeric molecules have corrupted. Pacybara detects recombinant clones and reduces false positive indel calls (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries).

Action D: Apply foldback-specific filtering. This is appropriate when foldback artifacts are detected. Foldback artifacts produce characteristic orientation patterns that can be recognized by alignment-based tools. Filtering should be applied before variant calling to prevent false structural variant calls (Detecting Foldback Artifacts in Long-reads).

Decision Point 5: Validate the Management Outcome

After applying the selected management action, validate that the outcome is correct. This validation step is often skipped but is essential for confirming that the management strategy achieved its goal.

For structural variant calling. Compare the number of structural variant calls before and after chimera management. A substantial reduction in calls, particularly inversions, suggests that chimeric reads were contributing false positives. The finding that restoring chimeras removed 97% of inversions on average in MDA-amplified samples provides a reference for expected reductions (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

For genome assembly. Check assembly contiguity and correctness metrics before and after chimera removal. Misassemblies caused by chimeric reads should be reduced, and the assembly graph should show fewer spurious connections.

For barcoded libraries. Verify that barcode-genotype associations are consistent after correction. Barcodes that were previously associated with multiple genotypes should now show a single consistent genotype (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries).

Implementing the Decision Framework in Practice

The decision framework can be implemented as a structured workflow with documented decision points. The Galaxy Training Network provides accessible tutorials for building reproducible analysis workflows that can incorporate these decision points. The nf-core Documentation describes community pipeline standards that support reproducible workflow configuration, which is useful for implementing the framework consistently across projects.

For researchers who need foundational skills in workflow implementation, The Carpentries Lessons provide training in shell scripting, data management, and reproducible analysis practices. These skills are directly applicable to implementing a chimera management decision framework.

Recording the Decision Framework Outcomes

Each application of the decision framework should produce a record that documents:

• The primary downstream analysis goal
• The library preparation and amplification history
• The detection tools applied and their versions
• The number of reads flagged as chimeric or foldback artifacts
• The management action selected and the rationale
• The validation metrics before and after management

This record serves multiple purposes. It provides the information needed for methods sections in publications. It enables troubleshooting when problems arise in downstream analysis. It establishes a baseline for monitoring library preparation quality over time. And it allows comparison across experiments to identify systematic issues with particular protocols or sequencing conditions.

The NCBI Data Resources provide official repositories for sequence data and associated metadata. Including chimera detection and management metrics in the metadata for deposited datasets improves the utility of those datasets for secondary analysis and supports reproducibility.

Common Mistakes in Applying the Decision Framework

Applying restoration to unamplified samples. Restoration tools such as 3rd-ChimeraMiner are designed for amplified samples where chimera rates are high. Applying them to unamplified samples with low chimera rates may introduce unnecessary computational overhead and may not improve results.

Removing reads instead of restoring in amplified samples. In MDA-amplified samples where chimera rates can exceed 78%, removing all flagged reads would discard the majority of the data. Restoration is the appropriate action in this context because it preserves the non-chimeric portions of the reads (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

Skipping the validation step. Applying a management action without validating the outcome leaves uncertainty about whether the action was effective. The validation step is essential for confirming that false-positive variants were removed and that genuine variants were preserved.

Using a single detection tool for all artifact types. Different artifact types require different detection approaches. Foldback artifacts may be missed by general chimera detection tools, and barcode-associated recombination requires specialized detection at the barcode-genotype association step. The decision framework should map artifact types to appropriate detection tools.

Frequently Asked Questions

What exactly is a chimeric read in long-read sequencing?

A chimeric read is a single sequencing molecule that contains sequence segments originating from two or more distinct genomic locations. These segments are joined together in the read but are not contiguous in the source genome. Chimeric reads can arise from library preparation artifacts such as adapter ligation concatemers, from mis-priming during multiple displacement amplification, or from foldback structures that create inverted sequence patterns. They can also arise from genuine structural rearrangements in the sample genome, which is why distinguishing artifacts from biology requires careful analysis.

How do chimeric reads affect alignment accuracy?

Chimeric reads affect alignment accuracy because aligners must decide how to map a read whose segments originate from different genomic locations. The aligner may split the read into multiple alignments, each mapping to a different locus. These split alignments appear as primary and supplementary alignments in the SAM/BAM output. If the chimeric nature of the read is not recognized, the supplementary alignments can be misinterpreted as evidence of structural variation, leading to false-positive variant calls.

What are the main causes of chimeric reads in PacBio and Nanopore libraries?

The main causes are ligation artifacts during library preparation, mis-priming during multiple displacement amplification, foldback artifacts that create hairpin-like structures, and barcode-associated recombination in mutagenized libraries. Ligation artifacts occur when multiple DNA molecules are joined together during adapter attachment. MDA mis-priming produces chimeric sequences that accumulate with continued amplification (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). Foldback artifacts create inverted sequence patterns that can be mistaken for genuine structural variants (Detecting Foldback Artifacts in Long-reads). Barcode-associated recombination creates false associations between barcodes and genotypes in mutagenized library experiments (Pacybara: Accurate long-read sequencing for barcoded mutagenized allelic libraries).

How can I detect chimeric reads in my sequencing data?

Detection approaches include inspecting SAM/BAM alignment flags for supplementary alignments, using alignment-based detection tools such as Breakinator to flag chimeric and foldback artifacts (Detecting Foldback Artifacts in Long-reads), and using specialized pipelines such as 3rd-ChimeraMiner for MDA-amplified samples (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). For barcoded libraries, Pacybara can detect recombinant clones and correct barcode-genotype associations (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries). The choice of detection method depends on the library preparation history and the expected types of artifacts.

What is the difference between a chimeric read and a structural variant?

A chimeric read is a technical artifact in which sequence from different genomic locations is joined in a single molecule. A structural variant is a genuine genomic rearrangement such as a deletion, inversion, or translocation. Both can produce split alignments in long-read data, which is why they are easily confused. The distinction requires additional evidence, including the number of supporting reads, the characteristics of the junction, and consistency across samples. In MDA-amplified samples, the vast majority of recognized chimeric sequences are artifacts instead of genuine structural variants (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data).

How does multiple displacement amplification contribute to chimeric reads?

Multiple displacement amplification uses phi29 DNA polymerase to amplify DNA through a process that involves mis-priming events. These mis-priming events can join sequence from different genomic locations, creating chimeric molecules. The proportion of chimeric sequences accumulates with continued amplification, ranging from 42% to over 78% in one study (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). This means that longer amplification produces more chimeric molecules, and samples amplified by MDA require dedicated chimera detection and restoration before structural variant analysis.

What tools are available for managing chimeric reads in long-read data?

Several tools are available. 3rd-ChimeraMiner recognizes and restores chimeric structures in long-read sequencing data, particularly for MDA-amplified samples (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). Breakinator flags putative foldback artifacts and previously known chimeric artifacts using an alignment-based approach (Detecting Foldback Artifacts in Long-reads). Pacybara detects recombinant clones and reduces false positive indel calls in barcoded mutagenized libraries (Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries). These tools should be integrated into analysis pipelines before structural variant calling or genome assembly.

Should I remove chimeric reads or restore them to their original structures?

The choice depends on the analysis goal. For structural variant analysis, restoration is often preferable because it recycles supplementary alignments that would otherwise contribute false-positive calls. The 3rd-ChimeraMiner pipeline was constructed specifically to restore chimeras to their original structures, and this restoration removed 97% of inversions on average in MDA-amplified samples (Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data). For other analyses, such as variant calling in individual genes, removing chimeric reads may be simpler and sufficient. The decision should be based on the specific analysis and the expected impact of chimeric reads on the results.

Related Bioinformatics Guides

• Detecting Structural Variants with Long-Read Sequencing: Methods and Considerations
• Long-Read Sequencing Cost and Market: What to Expect
• Long-Read Sequencing for Isoform Quantification: Challenges and Solutions
• RNA Sequencing Data Analysis: From Raw Reads to Differential Expression
• Short-Read vs Long-Read Sequencing: Pros, Cons, and Selection Criteria

Related Clinical & Scientific Guides

• A Practical Guide to Detecting Antimicrobial Resistance Genes in Shotgun Metagenomic Data
• Computational Immunology: Modeling the Immune System
• How to Set Hard Filters for Germline Variant Calling: A Practical Guide to GATK Best Practices

References and Further Reading

• NCBI Data Resources. National Center for Biotechnology Information.
• EMBL-EBI Training. European Bioinformatics Institute.
• Bioconductor. Bioconductor Project.
• Galaxy Training Network. Galaxy Project.
• nf-core Documentation. nf-core.
• The Carpentries Lessons. The Carpentries.
• Exploration of whole genome amplification generated chimeric sequences in long-read sequencing data.. Briefings in bioinformatics, 2023.
• Detecting Foldback Artifacts in Long-reads.. bioRxiv : the preprint server for biology, 2025.
• Detecting foldback artifacts in long-reads.. BMC genomics, 2026.
• Pacybara: Accurate long-read sequencing for barcoded mutagenized allelic libraries.. bioRxiv : the preprint server for biology, 2023.
• Pacybara: accurate long-read sequencing for barcoded mutagenized allelic libraries.. Bioinformatics (Oxford, England), 2024.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.