Science Editor

"We still haven't sequenced any human genomes in their entirety. There's still about 5 percent missing – and it's not a trivial 5 percent, or at least, it may not be," Harvard University professor of genetics George Church told researchers at a recent talk at the National Institutes of Health.

But sequencing has now nevertheless yielded more than 2,200 genes, which are considered "highly predictive and medically actionable."

Medically actionable, he clarified, does not always mean that the therapy is directed at the gene; it can also mean the use of medicines that do not target the gene directly, or devices. But even in such cases, "it is better to know what you are dealing with than to just know the symptoms."

In his talk, Church described a research project, and its approach to data, that he hopes will break new ground in both gleaning information from genomics and, in a larger sense, the way people think about what is genetic and what is environmental: The Personal Genome Project, initiated in 2005.

In the 1990s, PGP, in one corner of the Internet, stood for pretty good privacy and described an encryption method for emails. The Personal Genome Project shares the acronym – but its volunteers, which number nearly 14,000 to date and are from 66 countries, share (albeit anonymously) their genome sequence with the research community as well as the general public.

They also share a variety of other personal biological and environmental information, measured as part of the project or captured from medical records, such as information about a person's microbiome.

Epigenetics are also measured "to the extent possible." While gene sequences can be had from any cell in the body, "these epigenetic measurements have to be done from the correct cell type." In fact, to be comprehensive, epigenetic measurements would have to be done from every cell – a level of completeness, Church noted dryly, that is not "embraceable by the volunteers." For a compromise, induced pluripotent stem cells can be used to create a variety of different tissue types from patients, which can then be used for epigenetic measurements.

Such personal information, Church said, is vital to an understanding of the effects of genetic variation, because all public impressions to the contrary, genetics are anything but deterministic. "The point of the genome is not to say to a patients, 'Here's your genetic destiny; get used to it.' . . . You can avoid genetic destiny by changing your environment."

An example is the gene for phenylalanine hydroxylase. Mutations in this gene means its carriers are unable to metabolize the amino acid phenylalanine, leading to severe mental retardation and, eventually, death. "You can consider it 100 percent genetic," Church said, because mutations will inevitably lead to disease if they are not diagnosed. But at the same time, with near-universal newborn screening for the disorder in developed nations, "now it is 100 percent environmental because you can avoid dietary phenylalanine. It's not a trivial exercise, but you can do it."

As such genes have been discovered, Church said, they have changed, in real ways, the meaning of genetic and environmental disease. And so, the personal genome project is trying to capture not just genetic variants, but also the circumstances where they matter to an individual.

"We're not going directly from your personal genome, with 3 million differences between you and the reference sequence, directly to traits. That's not the idea. The idea is that environment is a big component, and the revolution in sequencing is partly spilling over into environmental measures."

The personal genome project champions what Church termed the GET approach, summarized most succinctly by Genomes + Environments = Traits.

The challenges of such an approach do not end with collecting the data from volunteers; indeed, they only begin there. Another challenge is to tell individuals what their genetics mean for them personally.

"It is possible, in fact likely, that individuals will get a hold of their own data," Church said. And even though the volunteers donate their information for the good of science, they are, of course, curious about what it means for them personally.

The problem is that often the experts don't know what a specific genetic variant means to the individual either. Church described one individual with a hypertrophic cardiomyopathy allele that is described only once in literature. Church said that his team's first instinct was that a single case is "not worth getting too worked up about."

But they did decide it was worth asking around amongst clinicians, and did indeed find that the allele in question had been observed though not published. At least one person with hypertrophic cardiomyopathy was recommended for echocardiogram testing.

Another question is how many people should have access to such data, and more importantly, should be able to contribute to analyzing them on the project's site. Church argued in favor of broad access, citing individuals such as the father of X-linked adrenoleukodystrophy sufferer Lorenzo Odone and others who are insiders only in the sense that they had family members afflicted with the diseases they broke new ground on.

"We don't know who is going to make the breakthroughs," he said. "It could be real outsiders."