Association studies between gene variants (polymorphisms) and measured intermediate phenotypes, such as lipid/lipoprotein levels, or disease endpoints such as coronary artery disease, are commonplace in the literature. But have we learnt anything from the shortcomings in study design and analytical strategies that have resulted in much controversy in this field over the last few years? This review highlights some of these problems. Using the lipoprotein lipase gene as an example, we evaluate new approaches to identifying polymorphisms that will stand up to linkage disequilibrium/association studies with complex disorders in this post Human Genome Project age, and emphasize the importance of gene-environment interaction in assessing the impact of gene variants.