Comparing the effect of imputation reference panel composition in four distinct Latin American cohorts

Area of Science:

• Population Genetics
• Genomic Imputation
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) identify genetic risk factors but depend on accurate imputation.
• Existing imputation reference panels often lack diversity, primarily featuring European ancestry, leading to reduced imputation quality in underrepresented groups.
• Latin American populations are underrepresented, posing challenges for genetic studies.

Approach:

• Evaluated imputation quality on chromosomes 7 and X across four Latin American cohorts.
• Modified reference panels by adjusting the inclusion of Latin American, African, and European individuals.
• Assessed the impact of increasing Indigenous American admixture proportions within reference panels.

Key Points:

• Increasing Latin American representation in reference panels consistently improved imputation quality.
• Excluding Latin Americans from reference panels significantly degraded imputation quality across all cohorts.
• Differential impacts were observed when excluding European or African individuals, varying by cohort and chromosome.

Conclusions:

• Reference panel composition critically influences imputation accuracy for Latin American populations.
• Increasing Indigenous American admixture enhanced imputation quality, but effects varied across populations.
• Future reference panels require careful construction to ensure equitable imputation performance across diverse Latin American groups and genomic regions.