Improved Allele Frequencies in gnomAD through Local Ancestry Inference.

Area of Science:

• Genomics
• Population Genetics
• Bioinformatics

Background:

• The Genome Aggregation Database (gnomAD) is a key resource for allele frequencies in genomic research and clinical genetics.
• Existing gnomAD data aggregates individuals into broad genetic ancestry groups, potentially masking important variations within admixed populations.

Purpose of the Study:

• To apply local ancestry inference (LAI) to identify ancestry-specific allele frequencies in admixed populations.
• To assess the impact of LAI-derived frequencies on variant interpretation and clinical significance.

Main Methods:

• Local ancestry inference (LAI) was applied to over 27 million variants in Admixed American and African/African American cohorts.
• Ancestry-specific allele frequencies were calculated for these admixed groups.

Main Results:

• A significant proportion of variants (78.5% and 85.1%) showed at least a twofold difference in ancestry-specific frequencies.
• Incorporating LAI data would increase the gnomAD-wide maximum frequency for 81.49% of variants, impacting clinical interpretation.
• Clinically relevant frequency differences masked by aggregate data were revealed.

Conclusions:

• LAI provides higher resolution allele frequency data for admixed populations compared to traditional methods.
• LAI-informed allele frequencies can refine variant classification, potentially moving variants from Uncertain Significance to Benign or Likely Benign.