Some of the hype that accompanied the first publications of the human genome sequence in 2001 may have worn off. But these are still heady times for genomics, as demonstrated this week by the release of a greatly enhanced haplotype map or HapMap, which describes the most common forms of human genetic variation (see page 851).

The map builds on an earlier version published in 2005. It may, for example, shed some light on aspects of the genome that help to account for certain differences between people of different geographical origins (see page 762). There have been plenty of other research findings this year that demonstrate the power of genomics to deliver clues that could yield better medicine, including studies based on the HapMap that have uncovered lists of multiple genes that may be associated with the risk of developing specific diseases.

But there remain relatively few examples where this has led to better treatment options for patients and doctors. The difficulties of selecting relevant gene and protein markers, and then developing them into marketable tests that doctors will use, remain formidable (see page 770). And for patients, doctors and even some geneticists, there is growing frustration at the lack of clarity in some research findings, the difficulty in discerning which findings are of medical value, and the slow pace at which fuller knowledge of the links between genetics and disease is actually providing better diagnosis and treatment options.

On page 773 of this issue, Nature tells the story of Hugh Rienhoff, a trained geneticist and biotechnology entrepreneur, whose daughter was born with a collection of congenital defects. He has taken it upon himself to try to find out what the genetic cause might be — actually buying lab equipment and having her genes sequenced himself. He has even posted information about her condition, his theories as to what's causing it, and parts of her genetic sequence on the Internet.

Given the sharply falling costs of equipment and the wealth of information that is publicly available, we are getting to the point at which almost anyone with access to the Internet can do this. If that sounds a little scary, then perhaps it ought to. Scientists and patient advocates have always enjoyed a delicate relationship. Researchers are not prone to welcome what they may see as the intrusion of the public in the laboratory. And there is every chance that some people in Rienhoff's position will waste money pursuing dead ends. On the other hand, as more people begin to take an interest in rare or undiscovered disorders, more useful information is likely to be unearthed about both their genetics and their treatment.

Scientists and patient advocates have always enjoyed a delicate relationship.

But this means that clinical geneticists will have to revise the professional and ethical framework for collaborating with patients and their advocates, to help ensure that the information from the public provides clarity and not confusion. Some scientists are already thinking about how best to organize such information. On page 783, for example, Steven Brenner of the University of California, Berkeley, proposes a 'genome commons' to aggregate the accumulated knowledge on human genetic and phenotypic diversity.

At the same time, members of the public who choose to embrace a do-it-yourself approach to science need to be aware that they should not abandon existing, rational treatment options. And they should know that the fruits of their labours will rarely include the cast-iron answers that they may be seeking. For, as is so often the case in science, the most likely result of their efforts will be yet more unanswered questions for others to probe.