Sir

Your News Feature on the HapMap project (Nature 425, 758–759; 2003) characterizes the Human Genome Diversity Project (HGDP) as an “ill-fated exercise” that “never got off the ground”. This is not accurate. The birth of the HGDP required time and perseverance because of misunderstandings during its development, which led to a 1997 review by the US National Research Council. This reviewing committee was persuaded of the scientific merit of human genome diversity projects, while also recognizing the ethical and legal challenges implicit in studies of human variation.

Their faith in the HGDP has been rewarded. Since April 2002 a collection of more than 1,000 DNA samples from 51 populations representing most of the world's genome variation has been available to non-profit research laboratories through a collaboration between the HGDP and the Fondation Jean Dausset-CEPH in Paris (see http://www.cephb.fr/HGDP-CEPH-Panel).

All samples used for this resource were collected with proper informed consent; the privacy of the persons volunteering these samples remains protected. Since 1997, the HGDP has developed and promoted the best legal practices available for studying indigenous populations.

Within months of its public release, the HGDP–CEPH resource became the basis of the most comprehensive description of the genetic structure of human populations ever undertaken (N. A. Rosenberg et al. Science 298, 2381–2385; 2002). This publication, which was not focused on any clinical outcome, was selected as the best biomedical publication of 2002 by The Lancet.

An enlarged HGDP collection would enhance research on the history and structure of human populations from all parts of the world, allowing researchers to characterize the global variation in any genetic sequence of interest. This work is useful to both evolutionary and medical studies. Identifying the genetic bases of complex diseases (the focus of the HapMap project) and of variation in response to environmental exposures and to drugs requires an understanding of genomic variation among unaffected individuals (controls) of similar ancestry to affected individuals.

Defining appropriate control populations for modern genomic analysis is still in its infancy. It can be helped by continuing political and financial support for HGDP resources.