Plasma levels of HDL cholesterol (HDL-C) predict the risk of cardiovascular disease at the epidemiological level, but a direct causal role for HDL in cardiovascular disease remains controversial. Studies in animal models and humans with rare monogenic disorders link only particular HDL-associated mechanisms with causality, including those mechanisms related to particle functionality rather than cholesterol content. Mendelian randomization studies indicate that most genetic variants that affect a range of pathways that increase plasma HDL-C levels are not usually associated with reduced risk of cardiovascular disease, with some exceptions, such as cholesteryl ester transfer protein variants. Furthermore, only a fraction of HDL-C variation has been explained by known loci from genome-wide association studies (GWAS), suggesting the existence of additional pathways and targets. Systems genetics can enhance our understanding of the spectrum of HDL pathways, particularly those pathways that involve new and non-obvious GWAS loci. Bioinformatic approaches can also define new molecular interactions inferred from both large-scale genotypic data and RNA sequencing data to reveal biologically meaningful gene modules and networks governing HDL metabolism with direct relevance to disease end points. Targeting these newly recognized causal networks might inform the development of novel therapeutic strategies to reduce the risk of cardiovascular disease.