The Genome Variation Ontology (GVO) is an ontology for the systematic description of various genomic variations, including complex structural variations in genomes. With advances in the discovery and genotyping methods for genomic variations, it is expected that new types of complex structural variations (SVs) will be discovered. We have developed the Ge-nome Variation Ontology to annotate all types of genomic variations, which includes complex SVs. Terms on genomic variations from dbSNP, dbVar, gnomAD, SO, VariO, and HGVS were collected and clustered manu-ally to generate 47 concepts. GVO is available at http://genome-variation.org/resource/gvo In recent years, large-scale sequencing projects have been carried out, such as the 100,000 Genomes Project by Genomics England. And based on massive amounts of individual genome sequences, a number of genomic structural variations (SVs) have been reported. For example, gnomAD-SV [1] contains a large number of novel SVs discovered using an improved multialgorithm ensemble method against high-coverage WGS. In addition it is envisaged that the widespread use of long-read sequence technologies will generate an ever-increasing amount of information on SVs in the near future. As ontologies that are able to be used for annotation of genomic variations, the Sequence Ontology (SO) [2] and the Variation Ontology (VariO) [3] are available: SO is an ontology for annotation of sequence features, while VariO is an ontology for describing the effects, consequences and mechanisms of mutations in DNA, RNA and proteins. These include terms for canonical SVs, such as deletion and translocation. However, besides canonical SVs, gnomADSV, for example, also contains complex SVs classified into 11 subtypes, some of which are not available in the existing ontologies.
We have developed TogoVar database that integrates allele frequencies from Japanese
populations and providing annotations for variant interpretation [
We collected terms on genomic variation from gnomAD, dbSNP, dbVar, SO, VariO and HGVS and performed manual clustering on equivalent concepts among them. As a result, we obtained the 47 ontology classes corresponding to genomic variation types. Then we have organized these classes hierarchically under gvo:Variation to construct GVO. These concepts include several complex SVs, such as paired-duplication inversion and paired-deletion inversion, which are not available in the existing ontologies. GVO is available in the web site (http://genomevariation.org/resource/gvo) and BioPortal (https://bioportal.bioontology.org/ontologies/GVO). Figure 1 shows an example of the GVO classes.
We plan to use GVO with the FALDO ontology to convert genomic variations distributed in VCF format into RDF. Therefore we also introduce some properties needed for this in GVO.