- Structural variation refers to relatively large changes in the structure of a genome, including deletions, duplications, insertions, inversions, translocations, and copy-number variations. These changes can affect hundreds to millions of DNA base pairs and may alter gene dosage, disrupt coding sequences, change regulatory regions, or modify chromosome organization. Structural variants contribute to normal genetic diversity, population differences, adaptation, inherited disorders, cancer, and other complex diseases. Their analysis is an important part of genome research, clinical genetics, evolutionary biology, and modern bioinformatics.
- Unlike single-nucleotide variants and small insertions or deletions, structural variants often involve complex genomic rearrangements that are difficult to detect and represent consistently. A deletion removes a segment of DNA, while a duplication produces an additional copy of a genomic region. An insertion adds sequence at a genomic location, and an inversion reverses the orientation of a segment. Translocations involve rearrangements between different genomic regions or chromosomes. Copy-number variations describe gains or losses of DNA segments relative to a reference genome and commonly include deletions and duplications. More complex events may involve several breakpoints, inserted sequences, or combinations of different variant types.
- The Database of Genomic Structural Variation (dbVar) is an NCBI resource designed to archive and provide access to information about large-scale genomic variation. It stores structural-variant data submitted by research projects, laboratories, and other data producers, together with supporting study and sample information. dbVar is an archival resource rather than a clinical interpretation database: its primary purpose is to preserve submitted observations and make them accessible for comparison, reuse, and further analysis. It complements resources such as dbSNP, which primarily focuses on smaller sequence variants, and ClinVar, which collects interpretations of the clinical significance of genetic variants.
- dbVar organizes information around studies, samples, and structural-variant observations. Study records describe the project or investigation from which the data originated, while sample records provide information about the biological samples analyzed. Variant records describe the reported genomic changes, including their locations, types, sizes, and other available attributes. The database assigns stable identifiers to submitted records, allowing researchers to refer to specific studies and variant observations. These identifiers should be retained when downloading or integrating data because they help preserve links to the original submission and its supporting metadata.
- Structural-variant records can include genomic coordinates, chromosome or sequence accession, reference assembly, variant type, estimated breakpoints, size, supporting evidence, and relationships to genes or other genomic features. The exact information available depends on the submitting study and the methods used to identify the variants. Some records represent precisely resolved sequence changes, whereas others describe approximate intervals or regions with uncertain boundaries. A reported interval should therefore not automatically be treated as an exact molecular breakpoint. Researchers must also distinguish between the coordinates of a variant on a reference genome and the actual sequence of the rearranged chromosome.
- Reference genome assembly is particularly important when interpreting structural variation. Coordinates reported on GRCh37 may differ from those reported on GRCh38 or another assembly, and chromosome naming conventions can also vary between datasets. A structural variant must be interpreted against the correct reference assembly, with attention to chromosome identifiers, coordinate conventions, and whether interval endpoints are represented using zero-based or one-based numbering. Assembly conversion tools such as NCBI Remap and UCSC LiftOver can help translate coordinates, but conversions may fail or become ambiguous when genomic regions have changed, are repetitive, or are represented differently between assemblies. Original coordinates and assembly information should always be retained alongside converted coordinates.
- A major category of structural variation is copy-number variation (CNV), which refers to differences in the number of copies of a genomic segment among individuals. CNVs may affect individual genes, multiple genes, or large chromosomal regions. They can arise through mechanisms such as non-allelic homologous recombination, replication-based errors, and DNA repair processes. CNVs may be inherited or arise de novo, and their effects depend on genomic location, size, gene content, regulatory context, and inheritance pattern. Some are common and have little or no measurable effect on health, while others contribute to developmental disorders, genomic syndromes, cancer, and variation in complex traits.
- Several complementary databases and resources support structural-variation research. Database of Genomic Variants (DGV) is a widely used resource for structural variants observed in apparently healthy individuals and population studies. It is particularly useful when assessing whether a reported deletion or duplication has previously been observed in control or general-population datasets. However, a variant appearing in DGV should not automatically be considered benign, because frequency, genomic context, phenotype, penetrance, and study design all influence interpretation. DGV and dbVar have related but distinct purposes: dbVar archives submitted structural-variation studies and observations, whereas DGV provides a curated catalogue emphasizing variation reported in control populations.
- The European Variation Archive (EVA), maintained by EMBL-EBI, provides an archival framework for genetic variation data, including structural-variant submissions. It supports data deposition, accessioning, retrieval, and reuse, and can complement NCBI resources when researchers need to locate variation datasets submitted through different international channels. The International Genome Sample Resource (IGSR) and the 1000 Genomes Project provide population-scale genomic variation data that can help characterize structural variation across populations. Large sequencing initiatives, including the Human Pangenome Reference Consortium (HPRC), further expand the representation of genomic diversity by providing assemblies and variation resources that capture sequences and structural differences not adequately represented by a single linear reference genome.
- Other resources are useful for particular types of structural variation or research questions. gnomAD provides population-frequency information for several classes of genetic variation, including structural variants in relevant releases. DECIPHER supports the interpretation of genomic variants in the context of human phenotypes and clinical research. ClinVar contains clinical significance assertions for variants, including some copy-number and structural variants, while ClinGen develops expert guidance and resources for clinical variant interpretation. Cancer-focused resources such as COSMIC and the Catalogue of Somatic Mutations in Cancer provide information about somatic genomic alterations, including rearrangements and copy-number changes in tumors. These databases should be used according to their scope: an archival record, a population observation, a clinical assertion, and a cancer-associated event are not interchangeable forms of evidence.
- Structural variants can be detected using several experimental and computational approaches. Chromosomal microarray analysis, including array comparative genomic hybridization and SNP arrays, has historically been important for identifying large deletions and duplications. Short-read next-generation sequencing can detect structural variants through read-pair discordance, split reads, read-depth changes, and local assembly. Long-read sequencing technologies, including Oxford Nanopore and PacBio platforms, can improve the detection and characterization of complex rearrangements because longer reads may span repetitive regions and structural-variant breakpoints. Optical genome mapping can provide information about large-scale structural changes, while cytogenetic methods such as karyotyping and fluorescence in situ hybridization remain useful for selected chromosome-level abnormalities.
- Bioinformatics tools are used to call, filter, annotate, and compare structural variants from sequencing data. Common tools include Manta, DELLY, LUMPY, GRIDSS, Sniffles, cuteSV, and SVIM, with suitability depending on sequencing technology, variant type, and analysis design. CNV-focused tools include CNVnator, Canvas, and GATK gCNV. These tools may use different algorithms and produce different breakpoint estimates, so combining evidence from multiple callers can improve confidence, although it can also increase analytical complexity. Read depth, mapping quality, allele balance, breakpoint-spanning reads, local sequence complexity, and technical artifacts should be considered when evaluating candidate variants.
- The Variant Call Format (VCF) is commonly used to represent structural variants identified through sequencing. Structural-variant records may use symbolic alleles such as <DEL>, <DUP>, <INS>, and <INV>, together with fields such as SVTYPE, END, SVLEN, and breakpoint-related annotations. More complex events may require additional fields or breakend notation. The BED format is frequently used for genomic intervals, while formats such as BEDPE represent paired genomic breakpoints. Researchers should not assume that every structural-variant dataset uses the same representation or coordinate conventions. Normalization, consistent reference assemblies, and careful handling of uncertain breakpoints are essential when comparing records from different sources.
- Retrieval and integration of structural-variation data can be performed through database websites, downloadable files, and programmatic interfaces. NCBI provides access to dbVar records and related resources through its web services and data-download facilities. Researchers can search by organism, study, variant type, genomic region, accession, or other available metadata. Bulk downloads are generally more suitable than repeated individual queries for large-scale analyses. Programmatic retrieval can be incorporated into Python or R workflows, allowing records to be filtered, converted, annotated, and combined with gene, phenotype, and population-frequency information. When using an API or downloadable release, researchers should record the retrieval date, database release or file version, query parameters, and reference assembly.
- A common application of structural-variation databases is the interpretation of candidate variants identified in genome sequencing studies. For example, a deletion detected in a patient may be compared with dbVar and DGV records to determine whether similar events have been reported previously. Population databases can help estimate how frequently the variant occurs, while gene and phenotype resources can identify potentially affected genes and relevant clinical features. Clinical interpretation requires additional assessment of technical validity, dosage sensitivity, inheritance, phenotype consistency, segregation, and established classification guidelines. A database match alone is not sufficient to establish pathogenicity.
- Structural-variation resources are also important in population genetics and evolutionary studies. Researchers can compare variant frequencies across populations, investigate genomic regions affected by recurrent rearrangements, and study the contribution of structural variation to adaptation and population history. In association studies, structural variants may be tested for relationships with diseases or quantitative traits. In cancer genomics, copy-number gains, deletions, and chromosomal rearrangements can help characterize tumor genomes, identify candidate driver events, and investigate mechanisms of genomic instability. In agricultural and conservation genomics, structural variation may contribute to economically important traits, adaptation, fertility, disease resistance, and genetic diversity.
- Despite their value, structural-variation databases have several limitations. Submitted datasets differ in sequencing technologies, sample selection, variant-calling algorithms, quality-control procedures, and reporting standards. Breakpoints may be uncertain, and the same biological event may be represented differently by different studies or software tools. Repetitive genomic regions and segmental duplications can complicate alignment and variant detection. Population representation is uneven, and some datasets contain limited information about ancestry, phenotype, or sample characteristics. The absence of a variant from a database does not demonstrate that it is rare or novel in all populations, while the presence of a variant does not establish that it is harmless or clinically relevant.
- Redundancy and inconsistent representation are additional challenges. A single structural event may appear in several submissions, sometimes with different coordinates or estimated sizes. Conversely, variants that look similar at the interval level may represent distinct molecular events. Researchers should compare genomic context, breakpoint evidence, affected sequence, sample information, and study provenance rather than relying only on overlapping coordinates. When integrating datasets, it is useful to preserve original records and identifiers while creating a standardized analysis representation. Any merging or deduplication procedure should be documented because different rules can change the number and characteristics of reported variants.
- Reproducible structural-variation analysis requires careful documentation of the reference genome assembly, annotation versions, variant-calling software and parameters, filtering criteria, coordinate conventions, database accessions, and retrieval dates. Quality-control measures should be reported, and uncertain or low-confidence calls should be distinguished from well-supported variants. Researchers should also consider whether a database is archival, population-based, clinical, or disease-specific before using its records as evidence. Combining dbVar with DGV, EVA, gnomAD, ClinVar, ClinGen, IGSR, and specialized disease or cancer resources can provide a more complete picture, provided that their different scopes and evidence types are respected.
- Overall, dbVar and related structural-variation resources provide essential infrastructure for storing, discovering, comparing, and interpreting large-scale genomic changes. They connect individual variant observations with the studies and samples from which they were derived and support research across genomics, population biology, clinical genetics, and evolutionary analysis. Their effective use depends on understanding structural-variant representations, reference assemblies, evidence quality, database scope, and the limitations of current detection technologies. Together with specialized population, clinical, and functional resources, these databases enable a more comprehensive understanding of genomic variation beyond single-nucleotide changes and small insertions or deletions.