Advances in DNA sequencing technologies have made it feasible to obtain near-complete genome sequences from thousands of individuals. Association mapping studies will immediately benefit from these developments: whole-genome sequencing of large GWAS datasets will not be practical for a while yet; in the meantime, we can impute a wide range of genetic variation from genomes that have already been sequenced. Most of the mutations discovered in these genomes will occur at low population frequencies, so it is important that imputation strategies be tailored to capture low-frequency variants. As sequencing projects produce larger and more diverse reference datasets, imputation-based GWAS will also face practical challenges like choosing appropriate reference panels and keeping computation tractable.
