Population structure at different minor allele frequency levels

Area of Science:

• Human population genetics
• Statistical genetics
• Genomic analysis

Background:

• Inferring population genetic structure is crucial for understanding human history and refining association tests.
• The minor allele frequency (MAF) of genetic variants influences detected population structures, particularly for rare variants in association studies.

Purpose of the Study:

• To investigate the hypothesis that differences in detected population genetic structure are due to a "scale" effect related to variant MAF.
• To explore how rare versus common variants reveal distinct population structures.

Main Methods:

• Utilized the Genetic Analysis Workshop 18 dataset comprising genotypes for 142 unrelated individuals.
• Analyzed single-nucleotide polymorphism (SNP) data, focusing on variants with varying minor allele frequencies.
• Employed kernel principal component analysis (KPCA) with varying kernel bandwidths to assess structural differences.

Main Results:

• Demonstrated that rare variants are primarily shared locally, reflecting smaller-scale population structures.
• Showed that common variants are spread over longer distances, indicating larger-scale population structures.
• Observed that changing the kernel bandwidth in KPCA yields results comparable to analyzing variants of different frequencies.

Conclusions:

• The scale of detected population genetic structure is directly influenced by the minor allele frequency of the analyzed variants.
• Understanding this scale effect is critical for accurate genetic association studies, especially when rare variants are involved.