Appreciating the contribution of human genome copy-number variation (CNV) to clinical phenotypes is one of the compelling genetics challenges of the coming years. It is increasingly possible to pursue such investigations as an extension of genome-wide association studies (GWAS), enabled by innovations in the design and analysis of SNP (single nucleotide polymorphism) arrays and by progress in determining the genomic locations and population-genetic properties of the CNVs that segregate in the human population. Extensions of GWAS to CNV have already resulted in discoveries of both de novo and inherited CNV that are associated with risk of common disease. This review will discuss new approaches, recent findings and the analytical challenges involved in expanding GWAS to appreciate the contribution of CNV to human phenotypes.