Admixture analysis of spontaneous hepatitis C virus clearance in individuals of African descent

Abstract

Hepatitis C virus (HCV) infects an estimated 3% of the global population with the majority of individuals (75–85%) failing to clear the virus without treatment, leading to chronic liver disease. Individuals of African-descent have lower rates of clearance compared to individuals of European-descent and this is not fully explained by social and environmental factors. This suggests that differences in genetic background may contribute to this difference in clinical outcome following HCV infection. Using 473 individuals and 792,721 SNPs from a genome-wide association study (GWAS), we estimated local African ancestry across the genome. Using admixture mapping and logistic regression we identified two regions of interest associated with spontaneous clearance of HCV (15q24, 20p12). A genome-wide significant variant was identified on chromosome 15 at the imputed SNP, rs55817928 (P=6.18×10−8) between the genes SCAPER and RCN. Each additional copy of the African ancestral C allele is associated with 2.4 times the odds of spontaneous clearance. Conditional analysis using this SNP in the logistic regression model explained one-third of the local ancestry association. Additionally, signals of selection in this area suggest positive selection due to some ancestral pathogen or environmental pressure in African, but not in European populations.

Genotype imputation performance of three reference panels using African ancestry individuals

“…However, the relative performance of these newer imputation reference panels – with varying total sample size and number of African individuals – is unknown. In this study, we extend these prior imputation comparisons beyond the 1000G populations by evaluating genotype imputation performance using CAAPA, HRC, and 1000G reference panels in two independent populations of African Americans (Regan et al 2010; Duggal et al 2013; Wojcik et al 2014).…”

Genotype Imputation Performance of Three Reference Panels Using African Ancestry Individuals

“…This improvement in the number of variants imputed and the accuracy using population specific panels argues that LD patterns of ethnic-specific variants may not be captured by different ethnic groups with distinct ancestral genetic background [53], which might include haplotypes from irrelevant populations. We would expect a population specific panel like CAAPA, where ~50% are African Americans with African ancestry estimates of 76% or higher [17], would be optimal for imputing more rare and common variants with higher quality in a target sample with similar high proportion of African ancestry (African ancestry estimates averaging 79.5%) [23,24]. However, in our study, the higher number of individuals from populations from continental Africa contained in 1000G compared to CAAPA (N=504 vs. N=25) may have outweighed the larger number of total African ancestry individuals in CAAPA, and provided more information on parental haplotype diversity [54,55] improving the chances of a rare variant being effectively tagged by a characteristic haplotype in admixed individuals [44].…”

“…The current study includes a total of 3,747 African Americans participating in previous genome-wide association studies of spontaneous resolution of Hepatitis C viral infection (HCV cohort) and Chronic Obstructive Pulmonary Disease (COPD) from the COPDGene cohort, a multi-site study of heavy smokers [23][24][25]. Metrics of imputation performance, accuracy, genome coverage and annotation of variants were calculated in these two cohorts, separately.…”

“…African Americans from the HCV cohort: A genome-wide marker panel from 447 African Americans was used, as previously described [23,24]. Briefly, 2,401 African American individuals participating in a longitudinal cohort study or identified through blood repositories as having HCV infection (spontaneously resolved or persistent) were enrolled, and were genotyped as part of the HCV Genetics Consortium.…”

“…However, the relative performance of these newer imputation reference panelswith varying total sample size and number of African individuals -is unknown. In this study, we extend these prior imputation comparisons beyond the 1000G populations by evaluating genotype imputation performance using CAAPA, HRC, and 1000G reference panels in two independent populations of African Americans [23][24][25].…”