# Why Your Haplotype-Resolved Assembly Is Chimeric ## Quick Answer - Chimeric contigs in haplotype-resolved assemblies typically arise from collapsed repeats, uneven sequencing depth, or assembler parameter mismatches with your specific heterozygosity level. - Run a post-assembly alignment audit with long reads mapped back to your contigs to identify chimeric junctions before committing to downstream analysis. - No single assembler setting fixes all chimeric assemblies, so you must diagnose the specific cause using coverage and alignment patterns instead of applying generic corrections. ## Understanding Haplotype-Resolved Assembly and Chimerism Haplotype-resolved assembly aims to reconstruct each homologous chromosome copy separately, producing two distinct sequences for diploid organisms. This approach contrasts with collapsed assemblies that merge homologous regions into a single consensus sequence. The value of haplotype resolution lies in its ability to reveal allele-specific variation, structural differences between homologs, and phase information that matters for understanding gene expression, disease mechanisms, and evolutionary history. A chimeric contig is a sequence artifact where segments from different haplotypes or different genomic locations are incorrectly joined into one contiguous sequence. In haplotype-resolved assemblies, chimerism typically manifests as a contig that contains sequence from both haplotypes at a given locus, or that incorrectly joins sequences from distant genomic regions. This creates a false haplotype that does not exist in the biological sample. Chimeric contigs are particularly damaging because they corrupt the fundamental unit of your assembly. Every downstream analysis, including variant calling, structural variant detection, gene annotation, and comparative genomics, inherits the errors. A single chimeric junction can produce false structural variants, incorrect gene models, and misleading phylogenetic signals. The problem is widespread enough that major bioinformatics training resources address assembly quality assessment as a core skill. The [Galaxy Training Network](https://training.galaxyproject.org/) provides accessible workflows for genome assembly and quality evaluation, emphasizing that assembly validation is not optional but a required step in any serious genomics project. Similarly, the [EMBL-EBI Training](https://www.ebi.ac.uk/training) program includes assembly and analysis pathways that teach researchers how to evaluate the reliability of their assembled sequences. ## Why Haplotype-Resolved Assemblies Are Especially Vulnerable Haplotype-resolved assemblies face challenges that collapsed assemblies do not. The assembler must distinguish between allelic variation and sequencing errors, and it must decide whether two similar sequences represent two haplotypes or one haplotype with errors. This decision is made repeatedly across the genome, and each decision point is an opportunity for chimerism. ### High Heterozygosity Creates Ambiguity When heterozygosity is high, the two haplotypes differ substantially at many positions. This is generally good for haplotype separation because the assembler can distinguish the two copies. However, high heterozygosity also means that the sequence divergence between haplotypes can be mistaken for structural variation or assembly errors. The assembler may attempt to reconcile the two haplotypes into one sequence, creating a chimeric contig that switches between haplotypes at points of high divergence. ### Repetitive Regions Confuse Path Selection Repetitive sequences, including transposable elements, segmental duplications, and satellite repeats, create ambiguous paths in the assembly graph. When a read spans a repeat, the assembler cannot determine which copy of the repeat the read belongs to. In haplotype-resolved assembly, this ambiguity is compounded because each haplotype may have a different repeat structure. The assembler may join sequences from different haplotypes that share the same repeat, creating a chimeric contig. ### Sequencing Depth and Error Profiles The depth of sequencing coverage directly affects the assembler's ability to distinguish haplotypes. Low coverage means fewer reads support each haplotype, making it harder to separate them. High coverage can create its own problems, including increased computational demands and the potential for the assembler to over-split or mis-join sequences. The error profile of the sequencing technology also matters. Long-read platforms have different error patterns than short-read platforms, and these patterns influence how the assembler handles ambiguous regions. ## Core Principles of Haplotype Assembly Understanding the principles of haplotype assembly helps you diagnose chimerism when it occurs. The assembler works with a graph structure where reads are nodes and overlaps are edges. In haplotype-resolved assembly, the graph is expected to contain two paths for each diploid region, one for each haplotype. The assembler must decide which path corresponds to which haplotype and where the paths diverge and converge. ### The Assembly Graph and Haplotype Paths The assembly graph is a representation of the sequence space defined by the reads. In a perfect world, the graph for a diploid genome would contain two parallel paths for each chromosome, with the paths diverging at heterozygous positions and converging at homozygous positions. The assembler's job is to traverse these paths and output contigs. Chimerism occurs when the assembler switches from one path to the other. This can happen when the graph is ambiguous, such as at a repeat or a region of low coverage. The assembler may also switch paths if it incorrectly determines that the two paths are actually one path with errors. ### Phasing and the Role of Linked Reads Phasing is the process of assigning variants to one haplotype or the other. Some assemblers use linked-read data, such as Hi-C or linked-read sequencing, to phase the assembly. These data provide long-range information that helps the assembler determine which variants belong together on the same haplotype. When linked data are not available or are insufficient, the assembler must rely on the sequence data alone, which increases the risk of chimerism. ### The Role of the Assembler's Parameters Each assembler has parameters that control how it handles heterozygosity, repeats, and coverage. These parameters are not universal. They must be tuned to the specific characteristics of your sample and your sequencing data. The [nf-core documentation](https://nf-co.re/docs) provides guidance on configuring assembly pipelines, and it emphasizes that parameter choices should be based on your data, not on defaults. The [Bioconductor project](https://bioconductor.org/) also provides packages for assembly evaluation and parameter exploration, allowing you to test different settings and compare the results. ## Practical Workflow for Diagnosing Chimeric Contigs When you suspect chimerism in your haplotype-resolved assembly, you need a systematic workflow to identify the chimeric contigs, determine the cause, and fix the problem. The following workflow is based on standard bioinformatics practices and is supported by the training materials from the [Galaxy Training Network](https://training.galaxyproject.org/) and the [EMBL-EBI Training Resources](https://www.ebi.ac.uk/training). ### Step 1: Map Reads Back to the Assembly The first step is to map your original sequencing reads back to your assembled contigs. This is a critical quality check. If the assembly is correct, the reads should map consistently to the contigs, with no evidence of chimeric joins. If a contig is chimeric, you will see reads that map to the contig in a way that is inconsistent with the expected coverage and alignment pattern. Use a long-read aligner if you have long-read data, or a short-read aligner if you have short-read data. The alignment should be performed with parameters that allow for the expected error rates of your sequencing technology. The [Galaxy Training Network](https://training.galaxyproject.org/) provides tutorials on read mapping and alignment quality assessment that can guide you through this step. ### Step 2: Examine Coverage and Alignment Patterns After mapping, examine the coverage and alignment patterns across each contig. Look for regions where the coverage drops to zero or near zero, which may indicate a misassembly. Look for regions where the alignment quality is poor, which may indicate that the contig sequence does not match the reads. Look for regions where reads from one haplotype map to a contig that is supposed to represent the other haplotype. A common pattern for chimeric contigs is a sharp transition in coverage or alignment quality at the point where the chimeric join occurs. You may also see reads that span the join point but have a high number of mismatches or indels, indicating that the sequence is not consistent with the reads. ### Step 3: Check for Haplotype-Specific Patterns In a haplotype-resolved assembly, you should see two contigs for each diploid region, one for each haplotype. If you see only one contig for a region that should be diploid, the assembly may have collapsed the two haplotypes into one. If you see a contig that contains sequences from both haplotypes, the assembly is chimeric. To check for haplotype-specific patterns, you can compare the two contigs for a region. They should be highly similar but not identical, with differences at heterozygous positions. If the two contigs are identical, they may be duplicates of the same haplotype. If they are very different, they may be from different regions of the genome. ### Step 4: Use Assembly Evaluation Tools Several tools are available for evaluating assembly quality. These tools can provide metrics such as N50, L50, and the number of contigs, but they can also provide more detailed information about the assembly structure. The [Bioconductor project](https://bioconductor.org/) provides packages for assembly evaluation, and the [Galaxy Training Network](https://training.galaxyproject.org/) provides workflows that incorporate these tools. One useful approach is to use a tool that compares the assembly to a reference genome, if one is available. This can help you identify regions where the assembly is chimeric or collapsed. If no reference genome is available, you can use the read mapping approach described above. ### Step 5: Identify the Cause of the Chimerism Once you have identified a chimeric contig, you need to determine the cause. The cause will guide your fix. Common causes include: - High heterozygosity that the assembler could not handle - Repetitive regions that created ambiguous paths - Low coverage in specific regions - Incorrect parameter settings To determine the cause, examine the reads that map to the chimeric region. Look at the coverage, the error profile, and the sequence context. If the region is repetitive, you will see many reads that map to the region with low alignment quality. If the region has low coverage, you will see few reads. If the region is heterozygous, you will see reads that differ from the contig sequence. ## Common Failure Patterns and Their Causes Understanding the common failure patterns in haplotype-resolved assembly helps you diagnose problems quickly. The following patterns are frequently observed in practice and are supported by the training materials from the [Galaxy Training Network](https://training.galaxyproject.org/) and the [EMBI-EBI Training Resources](https://www.ebi.ac.uk/training). ### Pattern 1: Collapsed Haplotypes In this pattern, the assembler produces a single contig for a region that should have two haplotypes. The contig is a mosaic of the two haplotypes, with the sequence switching between them at various points. This is a form of chimerism that is often caused by low heterozygosity or by the assembler's inability to distinguish the two haplotypes. The coverage of the collapsed contig is typically higher than the coverage of the individual haplotypes, because reads from both haplotypes map to the same contig. The alignment pattern shows a mix of reads from both haplotypes, with some reads having mismatches to the contig sequence. ### Failure Pattern 2: Contig Switching In this pattern, the assembler produces a contig that switches from one haplotype to the other at a specific point. The contig is correct for the first part, then switches to the other haplotype for the second part. This is often caused by a region of low divergence between the haplotypes, where the assembler cannot determine which haplotype the reads belong to. The coverage of the contig is usually uniform, but the alignment pattern shows a transition at the switch point. Reads that span the switch point may have a high number of mismatches or indels. ### Failure Pattern 3: Repeat-Induced Chimerism In this pattern, the assembler joins sequences from different genomic regions that share a repetitive element. The contig contains a repeat that is not actually contiguous in the genome. This is often caused by the assembler's inability to resolve the repeat structure. The coverage of the contig may be higher than expected, because reads from multiple genomic regions map to the same repeat. The alignment pattern may show reads from different regions mapping to the same contig. ### Failure Pattern 4: Coverage-Drop Chimerism In this pattern, the assembler joins two sequences that are not contiguous because of a region of low coverage. The assembler may bridge the low-coverage region with a sequence that is not supported by reads. This is often caused by insufficient sequencing depth. The coverage of the contig drops sharply at the chimeric junction. The alignment pattern shows a gap in coverage at the junction. ## Practical Fixes for Chimeric Assemblies Once you have identified the cause of the chimerism, you can apply a fix. The fix will depend on the cause and the assembler you are using. The following fixes are based on standard bioinformatics practice and are supported by the training materials from the [Galaxy Training Network](https://training.galaxyproject.org/) and the [nf-core documentation](https://nf-co.re/docs). ### Fix 1: Adjust Heterozygosity Parameters If the chimerism is caused by high heterozygosity, you may need to adjust the assembler's heterozygosity parameters. Many assemblers have a parameter that controls the expected heterozygosity rate. If you set this parameter too low, the assembler may collapse haplotypes. If you set it too high, the assembler may over-split the assembly. To determine the correct heterozygosity for your sample, you can estimate it from the read data. You can also use a tool that estimates heterozygosity from the read data. The [Bioconductor project](https://bioconductor.org/) provides packages for this purpose. ### Fixing 2: Adjust Repeat Handling Parameters If the chimerism is caused by repetitive regions, you may need to adjust the assembler's repeat handling parameters. Many assemblers have parameters that control how they handle repeats, such as the minimum repeat length or the number of copies. You may need to increase the minimum repeat length to prevent the assembler from joining sequences that are not contiguous. ### Fixing 3: Adjust Coverage Parameters If the chimerism is caused by low coverage, you may need to increase the sequencing depth or adjust the assembler's coverage parameters. The assembler may have a parameter for the minimum coverage required to support a path. If you increase this parameter, the assembler may be more conservative and produce fewer chimeric contigs. ### Fixing 4: Use a Different Assembler If the chimerism persists after adjusting the parameters, you may need to use a different assembler. Different assemblers have different strengths and weaknesses. Some assemblers are better at handling high heterozygosity, while others are better at handling repeats. You may need to test multiple assemblers to find the one that works best for your data. ### Fixing 5: Post-Assembly Polishing After you have identified and fixed the chimeric contigs, you may need to polish the assembly. Polishing is the process of correcting errors in the assembly using the read data. The [Galaxy Training Network](https://training.galaxyproject.org/) provides tutorials on assembly polishing, and the [nf-core documentation](https://nf-co.re/docs) provides information on polishing pipelines. Polishing can correct small errors, such as single-nucleotide errors and small indels. It cannot fix large-scale chimerism. If you have a chimeric contig, you must fix the chimerism before polishing. ## Records and Measurements for Assembly Quality Keeping records of your assembly quality is essential for reproducibility and for diagnosing problems. The following measurements are important to track: ### Assembly Statistics - Number of contigs - N50 and L50 values - Total assembly length - Number of contigs that are chimeric ### Read Mapping Statistics - Percentage of reads that map to the assembly - Percentage of reads that map uniquely - Coverage distribution across the assembly ### Haplotype-Specific Statistics - Number of regions with two haplotypes - Number of regions with one haplotype - Number of regions with chimeric contigs ### Parameter Settings - The assembler version and parameters used - The date of the assembly - The version of the reference genome, if used These records are important for reproducibility. The [nf-core documentation](https://nf-co.re/docs) emphasizes the importance of reproducible workflows, and the [Carpentries lessons](https://carpentries.org/lessons) provide training on data management and reproducibility. ## Common Failure Patterns and Their Causes You can use the following table to quickly identify the likely cause of a chimeric contig based on the observed pattern. | Observed Pattern | Likely Cause | Recommended Fix | | --- | --- | --- | | Coverage drops sharply at a junction | Low coverage region | Increase sequencing depth or adjust coverage parameters | | Alignment quality is poor at a junction | High heterozygosity | Adjust heterozygosity parameters | | Reads from multiple regions map to the same contig | Repetitive region | Adjust repeat handling parameters | | Two contigs are identical for a region | Collapsed haplotypes | Adjust heterozygosity parameters | | Contig switches between haplotypes at a point | Low divergence between haplotypes | Adjust heterozygosity parameters or use a different assembler | ## Practical Implementation Steps The following steps provide a practical approach to diagnosing and fixing chimeric contigs in your haplotype-resolved assembly. ### Step 1: Document Your Assembly Parameters Before you start diagnosing, document the assembler version, the parameters you used, and the input data. This is important for reproducibility and for comparing different assemblies. The [nf-core documentation](https://nf-co.re/docs) provides guidance on documenting your workflow. ### Step 2: Map Your Reads to the Assembly Map your original sequencing reads to your assembled contigs. Use an appropriate aligner for your data type. The [Galaxy Training Network](https://training.galaxyproject.org/) provides tutorials on read mapping. ### Step 3: Examine Coverage and Alignment Patterns Examine the coverage and alignment patterns across the assembly. Look for regions where coverage drops, alignment quality is poor, or reads from different haplotypes map to the same contig. ### Step 4: Identify Chimeric Contigs Identify contigs that show signs of chimerism. Use the patterns described in the table above to diagnose the likely cause. ### Step 5: Apply the Appropriate Fix Apply the fix that is appropriate for the cause of the chimerism. This may involve adjusting parameters, using a different assembler, or increasing sequencing depth. ### Step 6: Reassemble and Reassess Reassemble the data with the adjusted parameters and reassess the assembly quality. Repeat the process until the chimerism is resolved. ### Step 7: Document the Final Assembly Document the final assembly, including the parameters used and the quality metrics. This is important for reproducibility and for reporting your results. ## Records and Measurements The following table provides a template for recording your assembly quality metrics. | Metric | Value | Notes | | --- | --- | --- | | Assembler version | | | | Input data | | | | Parameters used | | | | Number of contigs | | | | N50 | | | | L50 | | | | Total assembly length | | | | Number of chimeric contigs | | | | Percentage of reads mapped | | | | Percentage of reads mapped uniquely | | | | Coverage distribution | | | ## Common Failure Patterns and Their Escalation When you encounter a chimeric contig, you need to decide whether to fix it yourself or escalate the problem to a more experienced bioinformatician. The following criteria can help you make this decision. ### Escalate When - The chimerism is widespread and affects many contigs - The chimerism is caused by a complex repeat structure that you cannot resolve - The chimerism is caused by a problem with the sequencing data, such as contamination or low quality - You have tried multiple fixes and the chimerism persists ### Handle Yourself When - The chimerism is limited to a few contigs - The cause is clear and the fix is straightforward - You have the time and resources to iterate on the assembly ## Relevant Welfare and Safety Context While this article focuses on bioinformatics, it is important to consider the broader context of your research. If you are working with animal or plant data, the quality of your assembly can affect downstream analyses that have welfare and safety implications. For example, a chimeric assembly can lead to incorrect variant calls that affect breeding decisions or disease risk assessments. The [NCBI](https://www.ncbi.nlm.nih.gov/) provides resources for biological data that can help you understand the context of your data. The [EMBI-EBI Training Resources](https://www.ebi.ac.uk/training) provide training on the ethical and responsible use of biological data. ## Limitations of the Diagnostic Approach The diagnostic approach described in this article has limitations. It relies on the quality of your read data and the accuracy of your aligner. It also relies on your ability to interpret the alignment patterns. If you are not experienced in interpreting alignment patterns, you may miss subtle signs of chimerism. The approach also assumes that the chimerism is caused by a single factor. In reality, chimerism can be caused by multiple factors, and the fix may require a combination of adjustments. ## Professional Escalation Criteria If you are unable to resolve the chimerism after multiple attempts, you should escalate the problem to a more specialized bioinformaticist. You should also escalate if you suspect that the chimerism is caused by a problem in the sequencing data, such as contamination or low quality. ## A Decision Framework for Triage, Repair, and Reassembly of Chimeric Haplotype-Resolved Contigs When you identify chimeric contigs in a haplotype-resolved assembly, the immediate question is also what caused the problem but also which response is most efficient for your project timeline, compute budget, and downstream analysis goals. A structured decision framework helps you avoid the common trap of repeatedly adjusting parameters without a clear hypothesis, which wastes compute time and complicates reproducibility. This section provides a practical triage system that classifies chimeric contigs by severity, assigns an appropriate repair pathway, and defines measurable criteria for deciding when to repair a single contig, when to reassemble a subset of the data, and when to restart the full assembly. ### The Three-Tier Triage System The triage system sorts chimeric contigs into three tiers based on the number of affected contigs, the genomic context of the chimeric junctions, and the impact on downstream analysis. This system assumes you have already completed the read-mapping audit described in the main workflow and have a list of candidate chimeric contigs with their junction coordinates. **Tier 1: Isolated Chimerism** Tier 1 applies when fewer than five contigs show chimeric junctions and the junctions are confined to a single genomic region or a small number of unrelated regions. The affected contigs are typically long, and the chimeric junction is a single point where coverage or alignment quality changes sharply. This pattern usually indicates a local assembly error, such as a mis-join at a repetitive element or a low-coverage gap that the assembler bridged incorrectly. The appropriate response for Tier 1 is targeted repair. You can attempt to break the chimeric contig at the junction and extend the two resulting fragments using the original read data. This approach preserves the rest of your assembly and avoids the computational cost of a full reassembly. The [Galaxy Training Network](https://training.galaxyproject.org/) provides tutorials on manipulating assembly graphs and extracting sub-sequences that are useful for this targeted repair approach. **Tier 2: Regional Chimerism** Tier 2 applies when chimeric contigs are concentrated in a specific genomic region, such as a large segmental duplication, a centromeric region, or a highly heterozygous locus. The affected contigs may number between five and twenty, and they often share a common cause, such as a complex repeat structure or a region of extreme heterozygosity that the assembler could not resolve. The appropriate response for Tier 2 is a targeted reassembly of the affected region. You can extract all reads that map to the chimeric contigs, along with their paired or linked reads, and reassemble this subset independently. This approach is more efficient than a full reassembly because it focuses the assembler's effort on the problematic region. The [nf-core documentation](https://nf-co.re/docs) provides guidance on configuring assembly pipelines for subset data, including how to adjust parameters for smaller input datasets. **Tier 3: Genome-Wide Chimerism** Tier 3 applies when chimeric contigs are distributed across the entire assembly, affecting many contigs in different genomic regions. This pattern usually indicates a systematic problem, such as an incorrect heterozygosity parameter, a mismatch between the assembler's expected error rate and the actual error profile of your sequencing data, or a fundamental issue with the input data quality. The appropriate response for Tier 3 is a full reassembly with revised parameters. You should not attempt to repair individual contigs because the underlying cause will continue to produce new chimeric junctions. The reassembly should be treated as a new assembly project, with careful documentation of the revised parameters and the rationale for each change. The [nf-core documentation](https://nf-co.re/docs) emphasizes the importance of reproducible workflow configuration, and the [Carpentries lessons](https://carpentries.org/lessons) provide foundational training on version control and data management that supports this documentation process. ### The Repair Decision Matrix Once you have assigned a tier, you need to decide between three repair strategies: break and extend, local reassembly, or full reassembly. The following matrix provides a structured comparison of these strategies based on the tier, the cause of the chimerism, and the resources available. | Triage Tier | Chimeric Contig Count | Typical Cause | Recommended Strategy | Compute Cost | Time Cost | Success Likelihood | | --- | --- | --- | --- | --- | --- | --- | | Tier 1 | 1 to 5 | Repeat junction or coverage drop | Break and extend | Low | Hours | High | | Tier 2 | 5 to 20 | Complex repeat or high heterozygosity | Local reassembly | Medium | Days | Medium | | Tier 3 | More than 20 | Systematic parameter or data issue | Full reassembly | High | Weeks | Variable | The success likelihood in this table is a general expectation, not a guarantee. Your specific data characteristics will influence the outcome. For example, a Tier 1 chimeric contig caused by a single low-coverage gap is highly likely to be repaired by breaking the contig and extending the fragments with additional reads. A Tier 2 chimeric contig caused by a complex repeat structure may require multiple iterations of local reassembly before the region is resolved. ### The Break and Extend Protocol The break and extend protocol is the most targeted repair strategy and is appropriate for Tier 1 chimerism. The protocol involves four steps that you can perform with standard bioinformatics tools. **Step 1: Identify the Exact Junction** Use the read alignment data to identify the exact position where the chimeric junction occurs. The junction is typically marked by a sharp drop in coverage, a cluster of mismatches, or a change in the alignment pattern. You should record the contig name and the junction coordinate in your assembly log. **Step 2: Extract the Flanking Sequences** Extract the sequence on both sides of the junction. The length of the flanking sequences should be sufficient to capture the unique sequence context on each side. A common approach is to extract 10 to 50 kilobases on each side, depending on the size of the contig and the complexity of the region. **Step 3: Map Reads to the Flanking Sequences** Map the original reads to the two flanking sequences separately. This step determines whether the reads support the two fragments as separate sequences or whether they support a connection between them. If the reads support the two fragments as separate sequences, the chimeric junction is confirmed. If the reads support a connection, the junction may be a true biological feature, and you should not break the contig. **Step 4: Extend the Fragments** If the reads support the two fragments as separate sequences, you can extend each fragment using the reads that map to it. This extension can be done using a local assembly tool or by manually adding reads to the fragment sequence. The goal is to produce two complete contigs that represent the two true haplotypes or the two true genomic regions. The [Galaxy Training Network](https://training.galaxyproject.org/) provides tutorials on sequence extraction and read mapping that support this protocol. The [Bioconductor project](https://bioconductor.org/) also provides packages for sequence manipulation and alignment that can be used to automate parts of this workflow. ### The Local Reassembly Protocol The local reassembly protocol is appropriate for Tier 2 chimerism, where multiple chimeric contigs are concentrated in a specific genomic region. This protocol is more complex than the break and extend protocol because it requires you to define the boundaries of the affected region and to reassemble the region independently. **Step 1: Define the Affected Region** Use the read alignment data to define the boundaries of the affected region. The region should include all chimeric contigs and a buffer of at least 100 kb on each side to ensure that the reassembly captures the full context of the region. **Step 2: Extract Reads for the Region** Extract all reads that map to the affected region, including reads that map partially to the region and reads that span the boundaries. You should also include reads that are unplaced or that map to other contigs but have a high likelihood of originating from the region, such as reads that share a k-mer with the region. **Step 3: Reassemble the Region** Reassemble the extracted reads using the same assembler you used for the full assembly, but with parameters adjusted for the specific characteristics of the region. For example, if the region is highly repetitive, you may need to increase the minimum repeat length. If the region is highly heterozygous, you may need to increase the heterozygosity parameter. **Step 4: Validate the New Assembly** Map the original reads back to the new assembly and examine the coverage and alignment patterns. The new assembly should show consistent coverage and alignment quality across the region, with no evidence of chimeric junctions. You should also compare the new assembly to the original assembly to ensure that the new assembly is an improvement. The [nf-core documentation](https://nf-co.re/docs) provides guidance on configuring assembly pipelines for subset of subsets, and the [Galaxy Training Network](https://training.galaxyproject.org/) provides tutorials on assembly validation that support this protocol. ### The Full Reassembly Protocol The full reassembly protocol is appropriate for Tier 3 chimerism, where the entire assembly is affected. This protocol is the most resource-intensive and should be undertaken only after you have exhausted the other options. **Step 1: Diagnose the Root Cause** Before starting a full reassembly, you must identify the root cause of the chimerism. This diagnosis should be based on the patterns you observed in the chimeric contigs, the read mapping data, and the assembly statistics. Common causes include incorrect heterozygosity parameters, incorrect repeat handling parameters, and incorrect coverage parameters. **Step 2: Revise the Parameters** Based on the diagnosis, revise the assembly parameters. The [nf-core documentation](https://nf-co.re/docs) provides guidance on parameter configuration for assembly pipelines, and the [Bioconductor project](https://bioconductor.org/) provides packages for parameter exploration and evaluation. **Step 3: Reassemble the Data** Reassemble the data using the revised parameters. This step may take several days or weeks, depending on the size of the genome and the computational resources available. **Step 4: Validate the New Assembly** After the reassembly is complete, validate the new assembly using the same read-mapping and alignment audit that you used to identify the chimeric contigs in the original assembly. The new assembly should have a lower number of chimeric contigs and improved quality metrics. **Step 5: Document the Changes** Document the changes you made to the parameters and the results of the validation. This documentation is essential for reproducibility and for future reference. The [Carpentries lessons](https://carpentries.org/lessons) provide training on data management and reproducibility that supports this step. ### A Record System for Chimeric Assembly Management A structured record system is essential for managing the repair and reassembly process. The record system should track the following information for each chimeric contig and for each repair attempt. **Contig-Level Records** For each chimeric contig, record the contig name, the length of the contig, the location of the chimeric junction, the coverage at the junction, the alignment quality at the junction, and the likely cause of the chimerism. This information is essential for diagnosing the cause and for deciding on the appropriate repair strategy. **Repair-Level Records** For each repair attempt, record the date, the repair strategy used, the parameters used, the computational resources used, and the outcome. The outcome should include the number of contigs produced, the quality metrics of the new contigs, and the number of chimeric junctions remaining. **Assembly-Level Records** For each assembly version, record the assembler version, the input data, the parameters used, the assembly statistics, and the number of chimeric contigs. This information is essential for comparing different assembly versions and for identifying the cause of the chimerism. The [nf-core documentation](https://nf-co.re/docs) provides guidance on how to structure reproducible workflows, and the [Carpentries lessons](https://carpentries.org/lessons) provide training on data management and version control that supports this record system. ### Common Failure Patterns in the Repair Process The repair process itself can fail. Understanding the common failure patterns helps you diagnose and correct problems in your repair workflow. **Failure Pattern 1: The Break and Extend Protocol Produces Fragmented Contigs** If the break and extend protocol produces contigs that are too short or that have poor quality, the cause is often that the flanking sequences were too short or that the reads did not provide sufficient coverage for extension. You can address this by increasing the length of the flanking sequences or by using a different extension strategy. **Failure Pattern 2: The Local Reassembly Produces the Same Chimeric Junctions** If the local reassembly produces the same chimeric junctions as the original assembly, the cause is likely that the parameters were not adjusted sufficiently or that the region is inherently difficult to assemble. You can address this by adjusting the parameters more aggressively or by using a different assembler. **Failure Pattern 3: The Full Reassembly Produces a New Set of Chimeric Junctions** If the full reassembly produces a new set of chimeric junctions, the cause is likely that the root cause of the chimerism was not correctly identified. You should return to the diagnosis step and re-examine the data. **Failure Pattern 4: The Repair Process Is Too Slow** If the repair process is too slow, the cause is likely that the computational resources are insufficient or that the parameters are too conservative. You can address this by increasing the computational resources or by adjusting the parameters to be more aggressive. ### Professional Escalation Criteria for the Repair Process The repair process can be complex, and you may need to escalate the problem to a more experienced bioinformaticist. The following criteria indicate when escalation is appropriate. **Escalate When** - The chimerism is widespread and affects more than twenty contigs - The chimerism is caused by a complex repeat structure that you cannot resolve - The chimerism is caused by a problem with the sequencing data, such as contamination or low quality - You have tried multiple repair strategies and the chimerism persists - The repair process is taking too long and you are concerned about the project timeline **Handle Yourself When** - The chimerism is limited to a few contigs - The cause is clear and the repair strategy is straightforward - You have the time and resources to iterate on the repair process ### The Role of the Decision Framework in Reproducibility The decision framework is also a practical tool for fixing chimeric assemblies but also a component of a reproducible research workflow. By documenting the tier of each chimeric contig, the repair strategy used, and the outcome of the repair, you create a record that allows other researchers to understand the decisions you made and to reproduce your results. The [nf-core documentation](https://nf-co.re/docs) emphasizes the importance of reproducible workflows, and the [Carpentries lessons](https://carpentries.org/lessons) provide training on data management and version control that supports this process. The framework also helps you avoid the common mistake of applying generic corrections to all chimeric contigs. By separating the contigs into tiers and applying the appropriate repair strategy, you can focus your resources on the contigs that are most likely to be repaired successfully and avoid wasting time on contigs that require a more aggressive approach. ### Limitations of the Decision Framework The decision framework has limitations. It relies on the accuracy of the read mapping and the quality of the assembly statistics. If the read mapping is inaccurate or the assembly statistics are misleading, the framework may lead you to the wrong conclusion. The framework also assumes that the chimerism is caused by a single factor. In reality, chimerism can be caused by multiple factors, and the repair may require a combination of strategies. The framework is also limited by the quality of the input data. If the sequencing data are of low quality or if the coverage is insufficient, the framework may not be able to resolve the chimerism. In these cases, you may need to generate additional sequencing data or to use a different sequencing technology. ### Practical Implementation Steps for the Decision Framework The following steps provide a practical approach to implementing the decision framework in your assembly project. **Step 1: Document the Chimeric Contigs** Document the chimeric contigs in your assembly, including the contig name, the junction location, the coverage at the junction, and the alignment quality at the junction. **Step 2: Assign a Tier to Each Chimeric Contig** Assign a tier to each chimeric contig based on the number of affected contigs and the genomic location of the junctions. **Step 3: Select a Repair Strategy** Select a repair strategy for each tier. Use the break and extend protocol for Tier 1, the local reassembly protocol for Tier 2, and the full reassembly protocol for Tier 3. **Step 4: Execute the Repair Strategy** Execute the repair strategy and document the outcome. **Step 5: Validate the Repair** Validate the repair by mapping the reads back to the new contigs and examining the coverage and alignment patterns. **Step 6: Document the Final Assembly** Document the final assembly, including the repair strategies used and the quality metrics of the final assembly. ### The Role of the Decision Framework in the Broader Assembly Workflow The decision framework is a component of the broader assembly workflow. It is used after the initial assembly and before the final validation. The framework is designed to be used in conjunction with the other steps in the workflow, including the read mapping, the coverage analysis, and the assembly evaluation. The framework is also designed to be used in conjunction with the training materials from the [Galaxy Training Network](https://training.galaxyproject.org/) and the [EMBL-EBI Training Resources](https://www.ebi.ac.uk/training). These resources provide the foundational knowledge and practical skills that are necessary to implement the framework effectively. ### The Framework and the Broader Context of Assembly Quality The decision framework is not a substitute for a comprehensive assembly quality assessment. It is a tool that helps you respond to the specific problem of chimeric contigs. The framework should be used in conjunction with other quality assessment tools and metrics, such as the N50, the L50, and the number of contigs. The framework is also not a substitute for a careful analysis of the biological context of your data. The chimeric contigs may be caused by a biological feature, such as a structural variant or a copy number variation, that is not an assembly error. In these cases, the framework may not be appropriate, and you should consult with a specialist in the biological context of your data. ### The Framework and the Use of External Resources The framework relies on the use of external resources, such as the [NCBI](https://www.ncbi.nlm.nih.gov/) and the [EMBL-EBI Training Resources](https://www.ebi.ac.uk/training). These resources provide the data and the training that are necessary to implement the framework effectively. The [NCBI](https://www.ncbi.nlm.nih.gov/) provides access to the reference genomes and the sequence data that are used in the validation step. The [EMBL-EBI Training Resources](https://www.ebi.ac.uk/training) provide the training that is necessary to understand the assembly and the repair process. The framework also relies on the use of the [Bioconductor project](https://bioconductor.org/) and the [nf-core documentation](https://nf-co.re/docs). These resources provide the software and the documentation that are necessary to implement the framework. The [Bioconductor project](https://bioconductor.org/) provides the packages for the assembly evaluation and the parameter exploration. The [nf-core documentation](https://nf-co.re/docs) provides the guidance for the configuration of the assembly pipelines. ### The Framework and the Reproducibility of the Assembly The framework is designed to support the reproducibility of the assembly process. The framework requires you to document the chimeric contigs, the repair strategies, and the outcomes of the repair. This documentation is essential for the reproducibility of the assembly and for the comparison of different assemblies. The framework also supports the reproducibility of the assembly by providing a structured approach to the repair process. The framework avoids the common mistake of applying generic corrections to all chimeric contigs. Instead, the framework provides a structured approach that is based on the specific characteristics of the chimeric contigs. ### The Framework and the Future of the Assembly The framework is designed to be flexible and adaptable to the future of the assembly. The framework is not tied to a specific assembler or a specific sequencing technology. The framework can be used with any assembler and any sequencing technology. The framework can also be used with the new assembly tools and the new sequencing technologies that are being developed. The framework is also designed to be used with the new assembly strategies that are being developed. The framework can be used with the new assembly strategies that are based on the long-read sequencing, the linked-read sequencing, and the Hi-C sequencing. The framework can also be used with the new assembly strategies that are based on the graph-based assembly and the pangenome assembly. ### The Framework and the Training of the Researchers The framework is designed to be used by the researchers who are working on the assembly. The framework is also designed to be used by the researchers who are learning the assembly. The framework provides a structured approach to the repair process that is easy to learn and to apply. The framework is also designed to be used by the researchers who are teaching the assembly. The framework provides a structured approach to the repair process that can be used in the teaching of the assembly. The framework can be used to teach the researchers how to diagnose the chimeric contigs and how to apply the appropriate repair strategy. ### The Framework and the Collaboration of the Research The framework is designed to be used by the researchers who are working on the assembly. The framework is also designed to be used by the researchers who are collaborating on the assembly. The framework provides a structured approach to the repair process that can be used in the collaboration of the assembly. The framework provides a common language for the researchers who are working on the assembly. The framework provides the terms and the concepts that can be used to describe the chimeric contigs and the repair strategies. The framework also provides the metrics and the records that can be used to compare the results of the different researchers. ### The Framework and the Communication of the Research The framework is designed to be used by the researchers who are communicating the results of the assembly. The framework provides a structured approach to the repair process that can be used in the communication of the assembly. The framework provides the records and the metrics that can be used to communicate the results of the assembly. The framework also provides the documentation that can be used to communicate the decisions that were made during the repair process. The framework also provides the documentation that can be used to communicate the limitations of the assembly and the repair process. ### The Framework and the Future of the Research The framework is designed to be used in the future of the research. The framework is designed to be used with the new assembly tools and the new sequencing technologies. The framework is also designed to be used with the new assembly strategies and the new analysis methods. The framework is designed to be used in the future of the research. The framework is designed to be used with the new assembly tools and the new sequencing technologies. The framework is also designed to be used with the new assembly strategies and the new analysis methods. The framework is designed to be used in the future of the research. The framework is designed to be used with the new assembly tools and the new sequencing technologies. The framework is also designed to be used with the new assembly strategies and the new analysis methods. ## Frequently Asked Questions ### What is a chimeric contig in a haplotype-resolved assembly? A chimeric contig is a sequence artifact where two segments from different haplotypes or different genomic regions are incorrectly joined into one contiguous sequence. This creates a false haplotype that does not exist in the actual sample. ### How do I detect chimeric contigs in my assembly? You can detect chimeric contigs by mapping your reads back to the assembly and examining the coverage and alignment patterns. Look for regions where coverage drops sharply, alignment quality is poor, or reads from different haplotypes map to the same contig. ### What causes chimeric contigs in haplotype-resolved assemblies? Chimeric contigs are caused by high heterozygosity, repetitive regions, low coverage, or incorrect assembler parameters. The specific cause depends on the characteristics of your data and the assembler you are using. ### Can I fix chimeric contigs by adjusting assembler parameters? Yes, you can often fix chimeric contigs by adjusting the assembler's heterozygosity, repeat handling, or coverage parameters. The specific adjustment depends on the cause of the chimerism. ### Should I use a different assembler if I have chimeric contigs? If you have tried adjusting the parameters and the chimerism persists, you may need to use a different assembler. Different assemblers have different strengths and weaknesses, and one may be better suited to your data. ### What is the role of post-assembly polishing in fixing chimeric contigs? Post-assembly polishing can correct small errors in the assembly, such as single-nucleotide polymorphisms and small indels. It cannot fix large-scale chimerism. You must fix the chimerism before polishing. ### How do I know if my assembly is good enough for downstream analysis? You can assess the quality of your assembly using metrics such as N50, L50, and the number of chimeric contigs. You can also compare your assembly to a reference genome if one is available. ### What should I do if I cannot resolve the chimerism in my assembly? If you cannot resolve the chimerism after multiple attempts, you should escalate the problem to a professional bioinformaticist. You should also consider whether the problem is caused by the sequencing data itself. ## Related Bioinformatics Guides - [Long-Read Metagenome Assembly: Overcoming Challenges with Nanopore and PacBio Data](/knowledge/bioinformatics/long-read-metagenome-assembly-overcoming-challenges-with-nanopore-and-pacbio-data) - [Evaluating Metagenomic Assembly Tools: A Benchmarking Framework for Short-Read and Long-Read Data](/knowledge/bioinformatics/evaluating-metagenomic-assembly-tools-a-benchmarking-framework-for-short-read-and-long-read-data) - [Long-Read Sequencing for De Novo Assembly of Complex Genomes: Case Studies and Best Practices](/knowledge/bioinformatics/long-read-sequencing-for-de-novo-assembly-of-complex-genomes-case-studies-and-best-practices) - [Long-Read Sequencing Cost and Market: What to Expect](/knowledge/bioinformatics/long-read-sequencing-cost-and-market-what-to-expect) - [Long-Read Sequencing for Isoform Quantification: Challenges and Solutions](/knowledge/bioinformatics/long-read-sequencing-for-isoform-quantification-challenges-and-solutions) ## References and Further Reading - [NCBI Data Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [EMBL-EBI Training](https://www.ebi.ac.uk/training). European Bioinformatics Institute. - [Bioconductor](https://bioconductor.org/). Bioconductor Project. - [Galaxy Training Network](https://training.galaxyproject.org/). Galaxy Project. - [nf-core Documentation](https://nf-co.re/docs). nf-core. - [The Carpentries Lessons](https://carpentries.org/lessons). The Carpentries. - [Hi-C sequencing data from frontal cortex of laboratory rats.](https://doi.org/10.1038/s41597-025-06173-4). 2025. - [Emergence of the Ug99 lineage of the wheat stem rust pathogen through somatic hybridisation.](https://doi.org/10.1038/s41467-019-12927-7). 2019. - [Cost-effective high-throughput single-haplotype iterative mapping and sequencing for complex genomic structures.](https://doi.org/10.1038/nprot.2018.019). 2018. - [A genomic glance through the fog of plasticity and diversification in Pocillopora.](https://doi.org/10.1038/s41598-017-06085-3). 2017. - [qc3C: Reference-free quality control for Hi-C sequencing data.](https://doi.org/10.1371/journal.pcbi.1008839). 2021. > This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.