Why Do Different Databases Show Different Allele Frequencies?
By Sequencing Team, The team of bioinformaticians, Genetic Health Coaches, and writers at Sequencing.
Short answer: Allele frequency databases such as gnomAD, ALFA, TOPMed and ClinVar include different people, use different testing methods and were built for different purposes, so the same variant can show different frequencies. Pick the database that best matches your question and use it consistently.
Not all genetic databases are built the same way, and that is why the same variant can have a different reported frequency depending on which resource you check.
The differences come down to three main factors: who was included in the study, how their DNA was tested, and what the database was designed to do in the first place.
What is an allele frequency?
An allele is one version of a DNA sequence at a given position. An allele frequency describes how often that version appears in a group of people, usually written as a decimal or percentage. A frequency of 0.01 means 1% of the copies of that position in the studied group carry the variant. Labs use these numbers when classifying variants, because a variant that is common in the general population is less likely to cause a rare disease.³
Who was included?
Some databases draw from large, diverse groups of people, while others focus on participants from studies with specific health conditions, such as heart or lung disease. A variant that is common in one group may be rare or not found among the people studied in another group, so the population behind the data shapes the frequency estimate significantly. This is also one reason polygenic risk scores can perform differently across ancestries.
How was the DNA tested?
Some databases use whole genome sequencing, some use exome sequencing (which focuses on protein coding regions), and others combine data from multiple technologies, including genotyping arrays. These methods differ in which parts of DNA they examine, how well they detect genetic changes, and how the results are checked for accuracy. Databases can also report positions on different reference genome builds, so make sure you are comparing the same variant.
What was the database designed for?
Some resources are built specifically to estimate population level variant frequencies, while others, like ClinVar, are primarily designed to collect assessments of whether genetic variants are linked to health conditions.² Frequencies shown in ClinVar are often imported from other sources rather than independently generated, which is why going to the original population database directly helps explain how the frequency was calculated and which population it represents.
Common databases at a glance
- gnomAD (Genome Aggregation Database): aggregates exome and genome sequencing data from large numbers of people across many populations, and is widely used as a general population reference.¹
- ExAC (Exome Aggregation Consortium): an earlier exome only resource that gnomAD grew out of. Older reports may still cite it.
- ALFA (Allele Frequency Aggregator): an NCBI resource that summarizes frequencies from research studies in the dbGaP database.
- TOPMed: a National Heart, Lung, and Blood Institute program with whole genome sequencing from participants in heart, lung, blood and sleep studies.
- ClinVar: a public archive of reports on whether variants are linked to health conditions, not a population frequency database.²
Because of these differences, it is important to be consistent about which database you rely on and to understand the strengths and limitations of each one when interpreting allele frequency data. If a frequency affects how you read a result, a genetic counselor can help put it in context.
Get Tested With Whole Genome Sequencing
Sequencing reads 100% of your DNA, including over 30,000 genes. Every Sequencing Kit Bundle includes 30x whole genome sequencing, Next Gen Disease Screening for more than 15,000 conditions, and Genome Explorer, where you can explore the variants in your own data.
Shop All Whole Genome Sequencing Kit Bundles
References
Frequently Asked Questions
An allele frequency is how often a particular version of a genetic variant appears in a group of people. It is usually shown as a decimal or percentage, such as 0.01, meaning 1% of the copies of that position in the studied group carry the variant.
Databases differ in who was included, how their DNA was tested, and what the database was built for. Different populations, sequencing methods and quality checks all change the frequency estimate.
ClinVar is mainly designed to collect assessments of whether variants are linked to health conditions. Frequencies shown there are often imported from other sources, so check the original population database to see how they were calculated.
Use the database that best matches the population and testing method you care about, and be consistent. Large population resources such as gnomAD are commonly used, but no single database fits every question.
A variant that is common in the general population is less likely to cause a rare disease, so labs use allele frequencies when classifying variants. Comparing frequencies from different databases without context can be misleading.