ISB Releases Kaviar, World’s Largest Public Catalog of Human Genomic Variation

A typical pair of human genomes are 99.9% identical. To understand what makes one person more prone to disease than another, it is the differences that are interesting. At ISB we have collected these differences from data sources covering nearly 72,000 individuals and integrated them into a catalog called Kaviar (Known Variants).

At the turn of the century a massive effort was made to determine the complete 3 billion nucleotide sequence of a single human genome (actually a composite of the genomes of several individuals). The result is called the reference genome. A typical human has 2 to 7 million differences from the reference genome, most of which are single nucleotide variants (SNVs). A scientist studying a disease may take the genomes of many people with that disease and see which SNVs they have in common. The scientist will then want to know, for each such SNV, how often it has been observed in the population at large. Kaviar provides a useful initial estimate.

Kaviar was first created four years ago by ISB Senior Research Scientist Gustavo Glusman and co-workers. At the time, only a small number of human genome sequences were publicly available. Glusman and co-worker Denise Mauldin continued to develop Kaviar and added new genomes as they became available. Then, over the past six months, Glusman and co-worker Terry Farrah made a concerted effort to add to Kaviar all publicly available whole genome and exome – the portion of the genome that encodes proteins – data sources. They also added the genomes of 780 unrelated individuals sequenced by ISB. The resulting database, along with an enhanced user-interface, was released January 14, 2015. This release of Kaviar contains 151 million different SNVs, some of them occurring in most genomes, some occurring in a small percentage, and the majority occurring quite rarely (e.g. <0.1%). Many of these rare SNVs are expected to be sequencing mistakes, and the Kaviar user must bear this in mind. Kaviar also catalogs 27 million small insertions, deletions, and substitutions.

Of the ~72,000 individuals covered, 12% are from whole genome sequence data sources such as the UK10K and the 1000 Genomes Projects, ISB’s 780, the Genomes of the Netherlands and the Wellderly Project. The remaining are from exome data sources, most notably The Exome Aggregation Consortium (ExAC). Anyone can look up variants in these sources directly, but it would be a tedious process. Kaviar allows anyone to look up any variant in a simple, fast web interface. Further, Kaviar calculates the overall frequency of each variant across all Kaviar data sources. Results for a list of 30 variants are returned in about 2 seconds. The entire Kaviar database can also be downloaded in Variant Call Format or VCF – commonly used for describing the SNVs in one genome or a collection of genomes – for fast and private local use.


Kaviar is programmed in Perl and offered free of charge as Open Source Software. Kaviar may be used online as a programmatic web service or downloaded for local use from The database is also provided as a downloadable VCF file.