The Genome Aggregation Database brings together sequencing data from large studies to estimate the frequency of genetic variants in human populations.
A variant that is too common to explain a rare disorder may be deprioritized, while a rare variant may merit further investigation. The database therefore supports filtering and hypothesis generation rather than making a diagnosis.
Coverage, ancestry composition, relatedness and study design all affect the estimates. Absence from a database is not proof that a variant is pathogenic, and presence is not proof that it is harmless.
STKEY note: use gnomAD alongside clinical databases, functional evidence and qualified interpretation.