Evaluating Genome Assembly Quality: Metrics and Tools

Genome assembly quality assessment is the process of verifying that a reconstructed genome sequence accurately represents the biological reality of the organism being studied. For students, researchers, analysts, and life-science professionals, understanding assembly quality is essential before any downstream analysis, because errors in the assembly propagate into every subsequent conclusion. This article explains the core metrics used to evaluate assemblies, describes the main tools available including Merqury, QUAST, and BUSCO, and provides a practical checklist for assessing whether an assembly is fit for its intended purpose.

Why Assembly Quality Assessment Matters

Genome sequencing technologies produce short fragments of DNA that must be computationally reconstructed into complete chromosomes. This reconstruction process, called assembly, is imperfect. The limitations of sequencing techniques have led to dozens of assembly algorithms, none of which is perfect, and no single method has emerged as a recognized benchmark for comparing assemblers [5]. Each assembler makes different tradeoffs, and the resulting assemblies can vary substantially in quality even when produced from the same input data.

The scientific value of assembled sequences depends on the ability to evaluate their completeness using a biologically meaningful approach [6]. A genome assembly is not an end product in itself. It serves as the foundation for gene annotation, comparative genomics, variant calling, evolutionary studies, and conservation planning. If the assembly contains errors, every analysis built on top of it inherits those errors. For example, a genome with missing genes will produce misleading results in studies of gene family evolution, and an assembly contaminated with DNA from another species will confound any sequence-based biological inference [10].

Public databases contain many assemblies of varying quality. When researchers download a genome from NCBI or another repository, they need tools to assess whether that assembly meets their standards. The National Center for Biotechnology Information provides access to genome assembly databases and associated metadata, but the responsibility for evaluating quality rests with the user [2]. Understanding the strengths and limitations of each quality metric is therefore a core skill for anyone working with genomic data.

Core Assembly Quality Metrics

Contiguity Metrics: N50 and Related Statistics

The most commonly reported assembly quality metric is N50. The N50 value represents the length of the shortest contig or scaffold such that the sum of all sequences of that length or longer equals at least 50 percent of the total assembly length. In practical terms, a higher N50 indicates a more contiguous assembly with fewer gaps. Related metrics include N90, which uses the 90 percent threshold, and L50, which reports the number of sequences needed to reach the 50 percent threshold.

Contiguity metrics are useful for comparing assemblies produced by different methods or from different sequencing platforms. For example, a high-quality genome assembly of the mushroom Lactarius hatsudake strain JH5 reported a scaffold N50 of 223.2 kb and an N90 of 54.5 kb, with the N50 length being 43.6 percent longer than that of a previously published assembly of the same species [8]. This comparison demonstrates how N50 can track improvements in assembly quality across versions.

However, contiguity metrics have serious limitations. Measures of contiguity are not strongly correlated with information about the biological completion or correctness of the assembly, and the commonly reported N50 metric can be misleading [9]. A genome can have a high N50 while containing substantial errors, or a low N50 while being biologically accurate. The N50 value says nothing about whether the assembled sequence matches the actual genome of the organism.

Research has reinforced this concern. A meta-analysis of 611 published articles on eukaryotic genomes that used BUSCO scores in addition to N50 found that although assemblies with relatively high contig and scaffold N50 values consistently had high BUSCO scores, a high BUSCO score could also be obtained from assemblies with a low N50 [12]. This finding confirms that N50 is not a perfect proxy for all measures of genome accuracy. Researchers should report additional assessment metrics alongside N50 instead of relying on contiguity alone [12].

Completeness Metrics: BUSCO

BUSCO, which stands for Benchmarking Universal Single-Copy Orthologs, provides a biologically meaningful measure of assembly completeness. The method is based on the concept of universal single-copy genes, which are genes expected to be present in essentially all species within a particular lineage [6]. BUSCO assesses genome assembly and annotation completeness using evolutionarily informed expectations of gene content [7].

The BUSCO approach works by searching the assembly for a set of conserved genes that should be present as single copies. The results classify each gene as complete, fragmented, or missing. A high BUSCO completeness score indicates that the assembly contains most of the expected gene content. For example, a chromosome-scale genome assembly of the plant Piptanthus nepalensis reported 99.3 percent BUSCO completeness for the genome and 99.2 percent for the predicted protein set [18]. Similarly, a chromosome-level assembly of the schizothoracine fish Gymnodiptychus pachycheilus achieved 98.40 percent BUSCO completeness [19].

BUSCO scores provide information that complements contiguity metrics. While N50 measures the physical structure of the assembly, BUSCO measures whether the assembly contains the expected biological content. Both dimensions matter for overall quality. A genome assembly of the burrowing sea anemone Paracondylactis sinensis reported a contig N50 of 8.70 Mb and a scaffold N50 of 9.41 Mb with 95.91 percent BUSCO completeness [20]. The combination of these metrics gives a more complete picture than either alone.

Base-Level Accuracy: Merqury and Quality Values

Merqury is a reference-free tool for assembly evaluation based on efficient k-mer set operations [23]. The method compares k-mers in a de novo assembly to those found in unassembled high-accuracy reads to estimate base-level accuracy and completeness [23]. This approach does not require a reference genome, making it valuable for evaluating assemblies of previously unsequenced species.

Merqury produces a quality value, often abbreviated as QV, which estimates the base-level accuracy of the assembly. Higher QV values indicate fewer errors. For example, a chromosome-scale genome assembly of Tigridiopalma magnifica reported a Merqury QV of 42.09 [15]. In another study, polishing an Oxford Nanopore assembly of the NA24385 human genome increased the Merqury QV from 28.7 to 32.9 [14]. These values provide a quantitative measure of how many errors remain in the assembly.

For trios, where sequence data from both parents and offspring are available, Merqury can also evaluate haplotype-specific accuracy, completeness, phase block continuity, and switch errors [23]. This makes the tool particularly useful for evaluating phased assemblies. Merqury also generates k-mer spectrum plots that provide visual representations of assembly quality [23].

Structural Metrics: LTR Assembly Index and Telomere Detection

Beyond base-level accuracy and gene content, structural features indicate assembly quality. The Long Terminal Repeat Assembly Index, abbreviated as LAI, measures the completeness of repetitive element structures in the assembly. A higher LAI value indicates more complete assembly of these complex regions. The Tigridiopalma magnifica assembly reported an LAI value of 16.01 [15].

Telomere detection provides another structural quality indicator. Telomeric repeats at chromosome ends indicate that the assembly has successfully reconstructed the natural chromosome termini. The Tigridiopalma magnifica assembly identified telomeric caps at both ends of all pseudochromosomes [15]. A near telomere-to-telomere assembly of the catfish Pterocryptis cochinchinensis reported telomeric repeats at multiple chromosome ends [16]. The presence of telomeres suggests that the assembly has captured the full extent of the chromosomes.

Contamination Assessment: BlobToolKit

Genome assemblies can contain DNA from sources other than the target organism. Contaminant DNA may be introduced during sample processing or co-extracted alongside the target DNA [10]. If insufficient care is taken during assembly, the final assembled genome may be a mixture of data from several species [10]. Such assemblies can confound sequence-based biological inference and, when deposited in public databases, may be included in downstream analyses by users unaware of underlying problems [10].

BlobToolKit is a software suite designed to help researchers identify and isolate non-target data in draft and publicly available genome assemblies [10]. The tool can process assembly, read, and analysis files for interactive exploration in a browser-based viewer [10]. BlobToolKit can be used during assembly to filter non-target DNA, helping researchers produce assemblies with high biological credibility [10]. The tool has been run on eukaryotic assemblies publicly available in the International Nucleotide Sequence Data Collaboration, with results available through a public instance of the viewer [10].

Tools for Assembly Quality Assessment

QUAST

QUAST, which stands for Quality Assessment Tool for Genome Assemblies, evaluates and compares genome assemblies [5]. The tool improves on leading assembly comparison software with new ideas and quality metrics [5]. QUAST can evaluate assemblies both with a reference genome and without a reference [5]. It produces many reports, summary tables, and plots to help scientists in their research and publications [5].

When a reference genome is available, QUAST can align the assembly to the reference and identify misassemblies, mismatches, and indels. When no reference is available, QUAST reports contiguity statistics and other assembly features. The tool was used to compare several genome assemblers on three datasets in the original study [5]. QUAST is particularly useful for comparing multiple assemblies of the same genome to select the best one.

BUSCO

The BUSCO tool suite assesses the completeness of genomes, gene sets, and transcriptomes using their gene content as a complementary method to common technical metrics [6]. The software is implemented in Python, with datasets available for download [7]. BUSCO provides a biologically meaningful approach to completeness assessment that goes beyond technical measures like N50 [7].

BUSCO requires a lineage-specific dataset of conserved genes. The user selects the appropriate lineage, such as aves_odb10 for birds or a general eukaryotic dataset. The tool then searches the assembly for these genes and classifies each as complete, fragmented, or missing. The results include both a summary score and detailed information about each gene. A chromosome-level genome assembly of the Greater Scaup Aythya marila reported that 97.0 percent of the highly conserved genes in aves_odb10 were present intact in the assembly [11].

Merqury

Merqury provides reference-free assembly evaluation based on k-mer set operations [23]. The tool compares k-mers in the assembly to those found in unassembled high-accuracy reads [23]. This approach estimates base-level accuracy and completeness without requiring a reference genome [23]. For trios, Merqury can evaluate haplotype-specific accuracy, completeness, phase block continuity, and switch errors [23].

Merqury is particularly valuable for evaluating assemblies of species without close relatives that could serve as references. The tool is fast and robust for assembly validation on both human and plant genomes [23]. The k-mer spectrum plots generated by Merqury provide visual representations that help researchers understand the distribution of k-mers in the assembly and reads.

Additional Tools

Several other tools address specific aspects of assembly quality assessment. SQUAT is a sequencing quality assessment tool for data quality assessments of genome assemblies [24]. EvalDNA uses machine learning for comprehensive evaluation of mammalian genome assembly quality [26]. MaGuS provides quality assessment and scaffolding of genome assemblies using Whole Genome Profiling data [27]. AIEdit is a machine learning-based polisher that operates alignment-free and generalizes across sequencing platforms [14].

The availability of multiple tools reflects the complexity of assembly quality assessment. No single tool captures every dimension of quality. Researchers should select tools based on the specific questions they need to answer about their assembly.

Practical Workflow for Assembly Quality Assessment

Step 1: Define the Intended Use of the Assembly

Before assessing assembly quality, determine what the assembly will be used for. Different applications have different quality requirements. A genome used for gene annotation needs high BUSCO completeness. A genome used for variant calling needs high base-level accuracy. A genome used for comparative genomics needs accurate chromosome structure. The intended use determines which metrics matter most.

Step 2: Run Contiguity Analysis

Calculate basic contiguity metrics including N50, N90, and L50 for both contigs and scaffolds. These metrics provide a quick overview of assembly structure. Tools like QUAST can generate these statistics automatically [5]. Record the number of contigs and scaffolds, the total assembly length, and the GC content.

Step 3: Assess Gene Content Completeness

Run BUSCO with the appropriate lineage dataset for the organism being assembled [6][7]. Record the percentage of complete, fragmented, and missing BUSCO genes. A high percentage of complete genes indicates that the assembly contains most of the expected gene content. A high percentage of missing genes may indicate assembly problems or an inappropriate lineage dataset.

Step 4: Evaluate Base-Level Accuracy

Run Merqury using the high-accuracy reads that were used for the assembly [23]. Record the quality value and examine the k-mer spectrum plot. A high QV indicates few base-level errors. The k-mer spectrum plot can reveal problems such as collapsed haplotypes or residual heterozygosity.

Step 5: Check for Contamination

Run BlobToolKit to identify potential contaminant sequences in the assembly [10]. Examine the taxonomic distribution of sequences and look for unexpected taxa. Remove or filter contaminant sequences if they are present. This step is particularly important for assemblies that will be deposited in public databases.

Step 6: Examine Structural Features

Check for telomeric repeats at chromosome ends and assess the LTR Assembly Index if repetitive elements are of interest. These structural features indicate whether the assembly has captured complete chromosomes. The presence of telomeres at both ends of all pseudochromosomes is a strong indicator of assembly quality [15].

Step 7: Document and Report All Metrics

Report all quality metrics together instead of relying on a single score. The meta-analysis of 611 published articles found that N50 alone cannot predict gene space completeness [12]. Researchers should report contiguity, completeness, and accuracy metrics together to provide a complete picture of assembly quality.

At a Glance: Assembly Quality Metrics and Tools

Records and Measurements for Assembly Quality

Maintaining detailed records of assembly quality assessment is essential for reproducibility and for supporting publication claims. The following measurements should be recorded for every assembly:

Assembly length in base pairs. This value should be compared to the estimated genome size from k-mer analysis. A chromosome-scale genome assembly of Citrus sinensis Pera IAC spanned approximately 620 Mb [13]. The genome size of Lactarius hatsudake was estimated to be 63.84 Mb based on k-mer frequency distribution, while the final assembled genome was 76.7 Mb [8]. Discrepancies between estimated and assembled genome size may indicate assembly errors or biological features such as high heterozygosity.

Contig N50 and scaffold N50 values. These values should be reported separately because they measure different levels of assembly structure. The Pterocryptis cochinchinensis assembly reported a contig N50 of 31.4 Mb with the assembly anchored onto 28 chromosomes [16]. The Piptanthus nepalensis assembly reported a contig N50 of 39.5 Mb and a scaffold N50 of 111.5 Mb [18].

BUSCO completeness percentages. Record the percentage of complete, fragmented, and missing BUSCO genes. The Gymnodiptychus pachycheilus assembly reported 98.40 percent BUSCO completeness [19]. The Paracondylactis sinensis assembly reported 95.91 percent completeness [20]. The Tigridiopalma magnifica assembly reported 95 percent completeness [15].

Merqury quality value. Record the QV for the assembly. The Tigridiopalma magnifica assembly reported a QV of 42.09 [15]. The AIEdit study reported that polishing increased the QV of an ONT assembly from 28.7 to 32.9 [14].

Percentage of sequences anchored to chromosomes. This value indicates how much of the assembly has been placed into chromosome-scale structures. The Pterorhinus courtoisi assembly anchored 92.32 percent of sequences onto 39 pseudochromosomes [17]. The Gymnodiptychus pachycheilus assembly anchored 95.52 percent onto 25 pseudochromosomes [19]. The Piptanthus nepalensis assembly anchored approximately 99.0 percent onto nine pseudochromosomes [18].

Number of predicted protein-coding genes. This value provides context for the assembly quality. The Tigridiopalma magnifica assembly predicted 43,291 protein-coding genes [15]. The Pterocryptis cochinchinensis assembly annotated 31,353 protein-coding genes [16]. The Paracondylactis sinensis assembly identified 19,420 protein-coding genes [20].

Common Failure Patterns in Assembly Quality

Overreliance on N50

The most common failure pattern is reporting N50 as the sole indicator of assembly quality. Research has shown that N50 is not strongly correlated with biological completion or correctness [9]. A high N50 can be achieved while the assembly contains substantial errors, and a low N50 assembly can still be biologically accurate. The meta-analysis of 611 published articles found that a high BUSCO score could be obtained from assemblies with a low N50 [12]. Researchers should report multiple metrics to avoid this failure pattern.

Ignoring Contamination

Assemblies can contain DNA from non-target species without obvious signs of problems. Contaminant DNA may be introduced during sample processing or co-extracted alongside the target DNA [10]. If insufficient care is taken during assembly, the final assembled genome may be a mixture of data from several species [10]. Such assemblies can confound sequence-based biological inference and may be included in downstream analyses by users unaware of underlying problems [10]. Screening for contamination with BlobToolKit should be a standard step in assembly quality assessment [10].

Using Inappropriate BUSCO Lineage

BUSCO requires selection of an appropriate lineage dataset. Using the wrong lineage can produce misleading completeness scores. The choice of lineage should reflect the evolutionary position of the organism being assembled. For example, a bird genome should use the aves_odb10 lineage, as was done for the Greater Scaup assembly [11]. Using a more distant lineage may result in many genes being classified as missing even if the assembly is complete.

Confusing Completeness with Accuracy

An assembly can be complete in terms of gene content while containing base-level errors. BUSCO completeness measures whether genes are present, but it does not measure whether the sequence of those genes is accurate. Merqury QV provides a separate measure of base-level accuracy [23]. Both metrics should be reported together to provide a complete picture of assembly quality.

Failing to Validate with Multiple Methods

The Citrus sinensis Pera IAC assembly used technical validation including BUSCO, LTR Assembly Index, Inspector, and Merqury analyses, all supporting the accuracy and completeness of the assembly [13]. This multi-method approach is the standard for high-quality assemblies. Relying on a single validation method increases the risk of undetected errors.

Limitations of Assembly Quality Metrics

N50 Limitations

The N50 metric has been criticized for being misleading when used as the primary measure of assembly quality [9]. Contiguity measures do not provide information about biological completion or correctness [9]. Multiple research groups have rejected the overuse of N50 and sought to develop more informative metrics [9]. Researchers should understand that N50 is a structural metric, not a biological one.

BUSCO Limitations

BUSCO has potential limitations that researchers should understand. The meta-analysis of 611 published articles suggested that variations in BUSCO scores among assemblies with poor N50 scores may be related to the number of introns in conserved eukaryotic genes [12]. BUSCO measures the presence of conserved genes, but it does not assess the accuracy of those genes or the completeness of non-conserved regions. The tool is most informative when used alongside other metrics.

Merqury Limitations

Merqury requires high-accuracy reads for k-mer comparison [23]. The quality of the reads directly affects the reliability of the quality value estimate. Merqury is reference-free, which is an advantage for evaluating assemblies of previously unsequenced species, but it cannot detect structural errors that do not affect k-mer content.

Reference-Based Assessment Limitations

When a reference genome is available, QUAST can evaluate assemblies against the reference [5]. However, this approach is only applicable to new assemblies of finished genomes [5]. The problem of evaluating assemblies of previously unsequenced species has not been adequately considered by reference-based methods [5]. Reference-free tools like Merqury address this gap [23].

Safety and Regulatory Context for Genome Assembly Data

Genome assembly data may be subject to data sharing policies and ethical considerations. The National Institutes of Health Genomic Data Sharing Policy governs the sharing of genomic data generated through NIH-funded research [3]. Researchers should be aware of the requirements for data deposition and sharing before beginning assembly projects.

The FAIR Guiding Principles provide a framework for making data findable, accessible, interoperable, and reusable [4]. Applying these principles to genome assemblies ensures that the data can be used by the broader research community. Assemblies deposited in public databases should include comprehensive metadata about the assembly methods and quality metrics.

The National Center for Biotechnology Information provides access to genome assembly databases and associated resources [2]. Researchers should deposit their assemblies in appropriate databases and include quality metrics in the associated records. The EMBL-EBI Training program offers educational resources for researchers working with genomic data [1].

Professional Escalation Criteria

Researchers should seek additional expertise or escalate concerns when they observe the following situations:

The BUSCO completeness score is substantially lower than expected for the organism being assembled. This may indicate assembly problems, inappropriate lineage selection, or biological features such as high heterozygosity that complicate assembly.

The Merqury QV is below the threshold required for the intended downstream analysis. Different applications require different levels of base-level accuracy. Variant calling and clinical applications require higher accuracy than comparative genomics.

BlobToolKit identifies unexpected taxa in the assembly. Contamination requires careful investigation to determine whether the contaminant sequences should be filtered or whether the assembly should be redone.

The assembled genome size differs substantially from the k-mer estimated genome size. This discrepancy may indicate collapsed haplotypes, over-assembly, or contamination.

The assembly will be used for clinical or regulatory purposes. These applications require the highest standards of assembly quality and may require additional validation beyond standard metrics.

Frequently Asked Questions

What is the difference between contig N50 and scaffold N50?

Contig N50 measures the contiguity of the assembled contigs before scaffolding, while scaffold N50 measures the contiguity after contigs have been ordered and oriented into larger scaffolds. Scaffold N50 is typically higher than contig N50 because scaffolding joins contigs using additional information such as Hi-C data. For example, the Piptanthus nepalensis assembly reported a contig N50 of 39.5 Mb and a scaffold N50 of 111.5 Mb [18]. Both values should be reported because they measure different levels of assembly structure.

Why is N50 not sufficient for assessing assembly quality?

N50 measures contiguity but does not measure biological completion or correctness [9]. Research has shown that contiguity measures are not strongly correlated with information about the biological quality of the assembly [9]. A meta-analysis of 611 published articles found that a high BUSCO score could be obtained from assemblies with a low N50 [12]. N50 should be reported alongside other metrics such as BUSCO completeness and Merqury QV.

How does BUSCO assess genome assembly completeness?

BUSCO uses evolutionarily informed expectations of gene content to assess assembly completeness [7]. The method is based on universal single-copy genes, which are genes expected to be present in essentially all species within a particular lineage [6]. BUSCO searches the assembly for these conserved genes and classifies each as complete, fragmented, or missing [6]. The results provide a biologically meaningful measure of completeness that complements technical metrics like N50 [6].

What is a Merqury quality value and how is it calculated?

Merqury quality value, abbreviated as QV, estimates the base-level accuracy of an assembly using k-mer set operations [23]. The tool compares k-mers in the assembly to those found in unassembled high-accuracy reads [23]. Higher QV values indicate fewer base-level errors. For example, the Tigridiopalma magnifica assembly reported a QV of 42.09 [15]. Polishing an ONT assembly increased the QV from 28.7 to 32.9 [14].

When should I use reference-free quality assessment tools?

Reference-free tools like Merqury should be used when no reference genome is available for the species being assembled [23]. Recent long-read assemblies often exceed the quality and completeness of available reference genomes, making validation against those references challenging [23]. Merqury provides a fast and robust method for assembly validation on both human and plant genomes [23].

How can I detect contamination in my genome assembly?

BlobToolKit is a software suite designed to help researchers identify and isolate non-target data in draft and publicly available genome assemblies [10]. The tool can process assembly, read, and analysis files for interactive exploration in a browser-based viewer [10]. BlobToolKit can be used during assembly to filter non-target DNA, helping researchers produce assemblies with high biological credibility [10].

What metrics should I report in a genome assembly paper?

Researchers should report multiple metrics to provide a complete picture of assembly quality. The meta-analysis of 611 published articles urged increased reporting of additional genome assessment metrics in addition to N50 [12]. Standard metrics include contig N50, scaffold N50, BUSCO completeness, Merqury QV, percentage of sequences anchored to chromosomes, and number of predicted protein-coding genes. The Citrus sinensis Pera IAC assembly reported BUSCO, LTR Assembly Index, Inspector, and Merqury analyses [13].

How do I choose the right BUSCO lineage dataset?

The BUSCO lineage dataset should reflect the evolutionary position of the organism being assembled. For example, a bird genome should use the aves_odb10 lineage, as was done for the Greater Scaup assembly [11]. Using a more distant lineage may result in many genes being classified as missing even if the assembly is complete. The BUSCO documentation provides guidance on selecting the appropriate lineage dataset [6][7].

Related Bioinformatics Guides

• Transcriptome Assembly Without a Reference Genome
• The Cancer Genome Atlas (TCGA): A Computational Overview
• Long-Read Genome Assembly and Polishing Strategies
• MicroRNA Target Prediction Tools
• The Human Genome Project: Computational Triumphs

References and Further Reading

• EMBL-EBI Training. European Bioinformatics Institute.
• NCBI Data Resources. National Center for Biotechnology Information.
• Genomic Data Sharing Policy. National Institutes of Health.
• The FAIR Guiding Principles. Scientific Data.
• QUAST: quality assessment tool for genome assemblies.. Bioinformatics (Oxford, England), 2013.
• BUSCO: Assessing Genome Assembly and Annotation Completeness.. Methods in molecular biology (Clifton, N.J.), 2019.
• BUSCO: assessing genome assembly and annotation completeness with single-copy orthologs.. Bioinformatics (Oxford, England), 2015.
• A high-quality genome assembly of Lactarius hatsudake strain JH5.. G3 (Bethesda, Md.), 2022.
• Toward a more holistic method of genome assembly assessment.. BMC bioinformatics, 2020.
• BlobToolKit - Interactive Quality Assessment of Genome Assemblies.. G3 (Bethesda, Md.), 2020.
• A high-quality chromosomal-level genome assembly of Greater Scaup (Aythya marila).. Scientific data, 2023.
• Assessing genome assembly quality prior to downstream analysis: N50 versus BUSCO.. Molecular ecology resources, 2021.
• Haplotype-resolved genome of Citrus × sinensis 'Pera IAC', the most widely cultivated sweet orange in Brazil.. 2026.
• AIEdit: Alignment-free genome assembly polisher trained on spaced seed match patterns.. 2026.
• A chromosome-scale genome assembly of Tigridiopalma magnifica.. 2026.
• A near telomere-to-telomere genome assembly of Pterocryptis cochinchinensis.. 2026.
• Chromosome-scale Genome Assembly of the Critically Endangered Blue-crowned Laughingthrush (Pterorhinus courtoisi, Leiothrichidae).. 2026.
• Chromosome-Level Genome Assembly and Annotation of Piptanthus nepalensis (Hook.) Sweet.. 2026.
• Chromosome-level genome assembly and annotation of the schizothoracine fish Gymnodiptychus pachycheilus.. 2026.
• Chromosome-scale genome of the burrowing sea anemone Paracondylactis sinensis.. 2026.
• Automatic Outlier Detection for Genome Assembly Quality Assessment. IEEE International Conference on e-Science, 2013.
• A Method for Reference-Free Genome Assembly Quality Assessment. 2013.
• Merqury: reference-free quality, completeness, and phasing assessment for genome assemblies. Genome Biology, 2020.
• SQUAT: A Sequencing Quality Assessment Tool for data quality assessments of genome assemblies. BMC Genomics, 2019.
• Identification of errors in draft genome assemblies at single-nucleotide resolution for quality assessment and improvement. Nature Communications, 2023.
• EvalDNA: a machine learning-based tool for the comprehensive evaluation of mammalian genome assembly quality. BMC Bioinformatics, 2021.
• MaGuS: A tool for quality assessment and scaffolding of genome assemblies with Whole Genome Profiling™ Data. BMC Bioinformatics, 2016.

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