Genome-wide association (GWA) studies have identified hundreds of genetic variants associated with various human complex traits and provided new insights to relevant biological pathways as well as potential drug targets. However, many potential challenges still remain for interpreting GWA results. For one, the large number of statistical tests performed require a stringent threshold for declaration of genome-wide significance. As a result, real associations may not reach this threshold and may become lost in the sea of random associations. Thus, additional information from human demographics, evolutionary history, or other biological or functional data needs to be considered to identify reproducible associations that may otherwise only reach modest statistical significance. Another challenge is that the successes of GWA studies have been largely achieved by studying European-derived populations, so that extending these studies into additional non-European populations is needed for translating clinically actionable findings into additional populations worldwide.

This dissertation is broadly divided into two parts to address the concerns above. First, we considered incorporating information from evolutionary conservation into association studies. We characterized the dosage sensitivity of ultra- and near-ultraconserved elements in the human genome, and then demonstrated that variations within these elements are enriched for nominal associations with some of the traits related to human reproductive and overall fitness. Second, to extend GWA studies into non-European populations, we described the first comprehensive GWA study of two anthropometric traits (BMI and height) in African-derived populations. We placed an upper limit to the effect sizes of variants influencing these traits in the African-derived genome and showed that, as a class, the effect sizes appear similar to that found in the European-derived genome. Finally, to facilitate the future designs of genetic studies in diverse populations, we introduced methods to rapidly assess the genetic ancestry in populations of unknown origin by genome-wide genotyping of pooled DNA samples. Taken together, these results suggest that in furtherance of human genetic mapping and medicine, innovative statistical and analytical frameworks will be needed to fully utilize the information from data that already exist and are yet to come.