HPDD, human protein domain database, is a non-redundant database composed of all domains that uniquely map to the human proteome. HPDD was created by an exhaustive search of all protein domains currently available within a database of human proteins followed by a selection of all non-overlapping domains found within each of these proteins. Currently, HPDD is comprised of 4,488 human protein domains. In addition, biologically relevant information such as mutations, post-translational modifications, domain interactions, and data from genome-wide association studies, have been extracted from several sources and mapped to all protein domains in HPDD. HPDD data can be accessed using a gene, protein, or domain layer, and its visualization tools provide a unique, domain-level view of protein variations and modifications within the human genome. This database could provide new insights about the molecular events leading to diseased and healthy organisms and will aid scientists in generating new hypotheses concerning the role of protein domains in key complex biological systems. HPDD is freely available at http://bioinf.umbc.edu/hpdd