Genotype imputation performance of three reference panels using African ancestry individuals

Area of Science:

• Genetics
• Genomics
• Population Genetics

Background:

• Genotype imputation enhances genome-wide association studies (GWAS) by estimating unobserved genotypes.
• Imputation is well-established for European ancestry but less so for African descent populations due to smaller reference panels.
• Available reference panels for African ancestry include 1000 Genomes (1000G), CAAPA, and HRC.

Purpose of the Study:

• To compare the performance of 1000G, CAAPA, and HRC reference panels for genotype imputation in African Americans.
• To evaluate imputation quality, accuracy, and variant coverage across different panels.

Main Methods:

• Compared imputation performance using three reference panels (1000G, CAAPA, HRC) on 3747 African Americans from HCV and COPDGene cohorts.
• Analyzed the number of imputed variants, imputation quality (R² > 0.5), and accuracy for rare, low-frequency, and common variants.
• Assessed variant overlap and exclusivity when merging imputed data from the three panels.

Main Results:

• 1000G imputed significantly more variants than CAAPA and HRC.
• Imputation quality and accuracy were high across all panels for African Americans, increasing variant availability.
• HRC showed the highest imputation quality for overlapping variants, followed by 1000G, then CAAPA, with quality increasing with minor allele frequency.

Conclusions:

• The 1000G, HRC, and CAAPA panels effectively improve genotype imputation for African Americans.
• These reference panels are complementary, and their combined use maximizes variant discovery.
• An integrated African reference panel is recommended to further enhance imputation performance.