Accurate variant interpretation depends on access to representative population frequency data. In VarSome v13.18, we’ve expanded our population frequency resources through the integration of KRIBB KOVA and ToMMo jMorp, which offer population allele frequency data from the Korean and Japanese populations, respectively. These datasets help laboratories interpret variants with greater confidence by offering ancestry-specific context, supporting more equitable genomic analysis for increasingly diverse populations.
Population frequency data is one of the most important sources of evidence in clinical variant interpretation. It helps laboratories distinguish between rare variants that may be disease-causing and common variants that are more likely to be benign.
However, population frequency databases, similarly to other genomic data sources, tend to contain a disproportionate amount of data from individuals of European descent1. As genomic testing becomes increasingly accessible and commonplace, this imbalance becomes more impactful; many genetic variants appear at higher frequencies within specific populations compared to what is reflected within larger databases. This means that variants that are in fact common in one population can appear to be artificially rare, reducing confidence in variant filtering and interpretation.
Not only does this impact variant interpretation for labs and clinics within underrepresented regions, it is increasingly becoming a global challenge. It is now estimated that over 3.7% of the world’s population are international migrants, living outside their country of birth, rising from 2.8% at the start of the 21st century2. This demographic shift highlights the growing need for more informative data to ensure a better understanding of how our genes influence our health.
There are many examples of genetic variants that are exceedingly rare in one population, yet are common and impactful in others.
The APOL1 gene, for example, contains variants that dramatically increase the risk of severe end-stage kidney disease and focal segmental glomerulosclerosis. However, these variants are so rare in those of European descent that studies within this population missed the association. In individuals of West African ancestry, these variants and the associated conditions were far more common, likely due to the variants playing a protective role against infection by Trypanosoma brucei3.
Another example is that of Brugada syndrome, which results in sudden cardiac death, and most frequently occurs in Asian populations. A number of genes are associated with this condition, and a wide range of variants had been previously classified as pathogenic. A 2019 reassessment of these pathogenic variants found that 14 of them were found at a minor allele frequency ≥ 0.001 in Asian populations specifically, suggesting that they are too common in some ancestries to be disease-causing4.
Every patient deserves to benefit from genomic medicine, and accurate variant interpretation requires reference data that reflects real diversity. To address this, VarSome v13.18 introduces two new population frequency databases, KRIBB KOVA and ToMMo jMorp, which offer population allele frequency data from the Korean and Japanese populations, respectively.
KRIBB KOVA (Korean Variant Archive)5, developed by the Korea Research Institute of Bioscience and Biotechnology (KRIBB), contains allele frequency data derived from healthy Korean individuals. It provides a population-specific reference for variants observed in patients of Korean ancestry.
ToMMo jMorp (Japanese Multi Omics Reference Panel)6, developed by the Tohoku Medical Megabank Organization (ToMMo), is a comprehensive Japanese reference database combining genomic and other omics data.
In 2025, a study published in Scientific Reports showcased the impact of leveraging population allele frequencies from Korean and Japanese cohorts, including from KOVA and jMorp, when identifying variants associated with hearing loss. The study assessed variants found in the Deafness Variants Database and, using data from East Asian populations, the authors reclassified several pathogenic or likely pathogenic variants as benign, likely benign, or of uncertain significance. Over 3,000 variants of uncertain significance were reclassified as benign, and several East Asian founder alleles were identified7.
Together, these datasets expand the population frequency evidence available within VarSome, giving laboratories access to more representative frequency information when analysing data from those of Korean or Japanese ancestry. They are also valuable for labs serving increasingly diverse populations, including diaspora communities and cosmopolitan cities, where ancestry-specific reference data can provide important context.
Variant interpretation depends on bringing together multiple lines of evidence beyond population frequency, including predicted functional impact, clinical evidence, inheritance patterns, and computational predictions. VarSome integrates these evidence sources into a single interpretation workflow, and by incorporating these new population frequency databases directly into the platform, users can now benefit from immediate access to ancestry-specific data alongside other relevant evidence without having to consult external resources independently.
This integration directly strengthens the clinical and research utility of VarSome, giving users richer context when assessing variant pathogenicity and improving confidence in classification decisions. This represents a meaningful step toward making VarSome the most complete and globally representative variant interpretation platform available.
References: