Understanding the Genetic Basis of extreme levels of High and Low LDL- Cholesterol in African Populations

Abstract

Cardiovascular diseases (CVDs) are a growing public health concern across Africa, with dyslipidaemia, particularly elevated low-density lipoprotein cholesterol (LDL-C), being a key modifiable risk factor. However, the genetics of lipid metabolism in African populations remain poorly understood due to limited representation in genomic studies. This thesis investigates both monogenic and polygenic contributions to extreme LDL-C levels in Sub- Saharan African populations, leveraging the Africa Wits-INDEPTH Partnership for Genomic Studies (AWI-Gen) cohort. To explore monogenic dyslipidaemia, we performed targeted next-generation sequencing (NGS) on 191 individuals with extreme high and low LDL-C levels from West, East, and South Africa, focusing on 16 genes implicated in LDL-C regulation, including LDLR, APOB, and PCSK9. This analysis identified 356 unique variants, including known pathogenic mutations and novel variants predicted to be functionally disruptive. While 28 variants were previously associated with phenotypic outcomes in ClinVar or HGMD, most were classified as variants of uncertain significance (VUS). Interpretation was complicated by the presence of multiple variants per individual, epistatic effects, and the absence of family-based cascade screening. Nevertheless, the findings highlight the complex genetic architecture associated with LDL-C levels in African populations and reinforce the need for functional studies and better population-specific variant classification. In the polygenic arm of the study, we evaluated the predictive performance of four published polygenic scores (PGS) for LDL-C, developed from European, African-American, and multi- ancestry cohorts, across 10,603 participants from the AWI-Gen study. The variance explained (R²) by each PGS ranged from 0.03 to 0.205, with the Graham and Zhang scores that were both developed in African–ancestry cohorts achieving the highest overall accuracy. Notably, regional differences in performance were observed, with PGS trained on African-American datasets performing better in South and East Africa than in West Africa, despite African- American having predominantly West African ancestral admixture. These results suggest that current PGS have limited and uneven predictive power in African populations, highlighting the need for ancestry-diverse discovery cohorts and region-specific calibration. v Together, these findings emphasize the importance of incorporating African genetic diversity into both monogenic and polygenic frameworks to improve CVD risk stratification and enable more equitable implementation of precision medicine on the continent.