How much do you really know about yourself? You may know that you have allergies, or that your eyes are blue, or that your family history is rife with diabetes. But do you know whether you're at an elevated risk for developing breast cancer, or which mutations have been handed down to you from your mother and father that could contribute to your risk of osteoporosis? More importantly, do you want to know?
For researchers, every genome sequenced and analyzed for every bit of information it can yield could lead to better understanding of a disease or a connection between genotype and phenotype.
For clinicians, delving into the specifics of patients' genomes could help in finally convincing them to quit smoking or start exercising. But as everyone has come to learn, the question is not, "How much can you learn about yourself?" but instead: "How much do you want to know?"
Researchers, clinicians, and bioethicists are asking themselves and their patients those very questions as the next evolution of genomics — personal -genomics — enters the medical and scientific scenes. As sequencing becomes less expensive, and the fabled $1,000 genome comes closer to being a reality, it will soon be just as cheap to sequence an entire genome as it is to test for one or two disease-causing genes.
But just because the technology is getting faster doesn't mean doctors and their patients are keeping up. The questions they pose are simple: What do you do with all that data? Is it really necessary to sequence an entire genome just to determine someone's risk for cancer or diabetes? What if patients do not want to know? Answering these, it turns out, is not so easy.
In addition, the bioinformatics and analysis side of genome sequencing has a lot of catching up to do as researchers struggle to analyze and interpret the massive amounts of data being churned out by sequencers daily.
In 2006, five years after the draft human genome was released, Harvard University's George Church said it was time to start thinking about sequencing diploid genomes from all sorts of people and create an open-access database to store all the information from those genomes for the wider scientific community to use. "If we really want to understand Homo sapiens, we have to know all sorts of stuff about them," says Duke University's Misha Angrist, who was a founding member of Church's Personal Genome Project and one of the first people to have his genome published online for public use. "Not only their genomes, but what's in their medical records? What did they eat for breakfast? What is their emotional state? What do they do for a living? What kind of vitamins did they take? What are they allergic to? Do they take recreational drugs?" In the beginning, the PGP was about finding people — 1,000 at first — who were willing to forgo privacy and confidentiality to make all their data freely accessible, he says, adding, "One of the great promises of the Human Genome Project was that it would be good for public health and I know that the PGP brain trust has the same hopes for the PGP."
Then in 2007, J. Craig Venter made history when he published the first personal genome — his own — in PLoS Biology. And two years later, Stanford University's Stephen Quake made headlines when he sequenced and annotated his own entire genome. He was able to determine his risk for obesity, type 2 diabetes, and coronary artery disease as well as which drugs he should avoid, and what kind of statin he would respond to best if he needed one.
Even as Church, Angrist, Quake, and their colleagues continue their work, whole-genome sequencing is becoming easier to undertake thanks to some rapidly advancing technology. Sequencing technology has been on a steep upward curve, a circumstance that some researchers say is directly responsible for the rise of personal genomics.
Angrist, who studies the inter-section of genomics and society and has helped pioneer the personal genomics field, says that the past few years have seen enormous leaps in sequencing technology, starting with the 454 platform — which was an improvement on Sanger sequencing — and then Solexa. (Those startup companies were purchased by Roche and Illumina, respectively.) These next-gen sequencers provided "maturation and viable competition" to the field, he says. Now, companies like Pacific Biosciences and Complete Genomics have almost made whole-genome sequencing "trivial," as has Illumina's HiSeq technology, and as the newer wave of sequencing technologies based on electron microscopy, microfluidics, and nanopores promise to, Angrist adds.
Stanford's Quake is one of the founders of Helicos BioSciences, which specializes in genetic analysis technologies, like its Helicos Genetic Analysis Platform — which was the first DNA sequencer to image individual DNA molecules. Quake used that single molecule sequencing technology, the Heli-Scope, when he sequenced his own genome.
In December, Quake and his team published a paper in Nature Biotechnology in which they used micro-fluidics as a tool to do haplotyping of a human genome. "One of the imperfections of all the genomes that have been published to date is that they lack haplotype information," Quake says. He and his team have developed a method to lyse single cells to separate the chromosomes and amplify them individually in order to determine which genes were maternally inherited and which were paternally inherited.
Also in December, Jay Shendure, Evan Eichler, and their colleagues at the University of Washington School of Medicine published a paper in Nature Biotechnology on the individual haplotype — which, they wrote, is "essential to the complete description and interpretation of genomes, genetic diversity, and genetic ancestry." They combined massively parallel sequencing and large-insert cloning to experimentally determine the haplotype-resolved genome of a South Asian individual.
Because haplotyping separates the chromosomes and differentiates between mutations that may have been handed down by one's -mother, father, or both, "you can imagine there are clinical consequences for that information," Quake says. Micro-fluidics technologies enable this kind of analysis because they allow researchers to perform very complex manipulations with a small amount of genetic material. Smaller sample sizes also reduce the risk of contamination. "That ends up being practically important for these amplifications," he adds.
Bioinformatics lags behind
Whereas sequencing is advancing at a rapid pace — "It's all moving faster than Moore's Law," Quake says — the bioinformatics side is having a hard time keeping up in the analysis of all the data being generated. Analyzing whole-genome sequences is expensive, hard to do, and lacks the flash and pizzazz of creating new sequencing technologies.
"The bottleneck is in interpretation," Angrist says. "Having slogged through my own genome, I can say that I don't know that I would have spent $20,000 of my own money to look at Excel spreadsheets and think of what build of the genome I need to be looking at and [asking,] 'Is this an error or a stop codon? Why aren't I dead? What's going on here?' I'm glad I did it, but I'm glad it was on someone else's nickel."
Mildred Cho at the Stanford University Center for Biomedical Ethics says that although researchers doing sequencing are gung-ho about personal genomics, researchers who are developing analytic methods say it is not quite ready for prime time. "Technically, you can obtain large amounts of -information on individuals, but the analysis part of that is still quite hampered, and labor intensive, and incomplete," she says. Many of the databases bioinformaticians would need to access in order to make a complete analysis are privately owned, overly expensive to get into, or otherwise inaccessible, Cho says. Public databases do exist, but they are scattered about, making it hard for anyone to get a complete picture of the data that is out there. And even after researchers do get access to the data they need, she adds, many find that the data is not always robust enough to identify clinically significant conditions — many variants have only weak -associations with disease, and some rare variants turn out not to be associated with disease at all, making the entire analysis a far cry from cut-and-dried.
Michelle McGowan, assistant professor of bioethics at Case Western Reserve University, is of a similar mind to Angrist and Cho. "Interpretation is really where the hang-up is going to be," she says. "The people power that's needed, as well as deciphering information that's never been available before is going to be the real rate-limiting factor."
And doctors are also concerned about having to interpret the information for patients when they themselves are still learning about it. "We're not at a point where we can say personal genomics is ready," she says. Furthermore, clinicians are also worried about how they would bill for these genetic tests — there is no available infrastructure to integrate this kind of sequencing or even large-scale mutation screening into most medical practices, McGowan adds.
Cho says that in conducting studies on personal genomics analysis, respondents to her surveys have indicated that making sense of the data, even for themselves, is still problematic. "It's so much information that the clinical community has recognized that they don't understand it enough to be able to explain it to patients, and I don't think even researchers know enough to be able to explain it as a whole," she says. "The way people are doing things now is to use genetics in a targeted way, but as it gets cheaper, it doesn't make sense to do one gene or a set of genes anymore. It will be just as cheap to do the whole genome."
Cheaper sequencing, it seems, is a gift that comes with strings attached.
A problem of ethics
Ethical conduct questions have also cropped up for personal genomics. Stanford's Cho says that the major problem with uncovering large amounts of data about people and sharing it with them is that they may not want to hear about the secrets their genomes contain. However, the person who handles the data — the doctor who ordered the test, or bioinformatician who analyzed the information — may find himself or herself in a position of holding data that may be of potential clinical significance and not being able to do anything about it. "That puts the person who analyzes the data in a potentially uncomfortable situation, and there may be some obligation to do something about it," Cho says. "And the person whose sample it is may not want it and may not have been asked about what they want."
Case Western's McGowan is conducting a study on how people see themselves and their identities after undergoing genetic testing. She is also concerned with issues related to genetic discrimination and privacy. As yet, she says, there are no guidelines for the physical storage and safekeeping of genetic information. "Can you trust that your genomic information is being held safely?" she asks. And what does that mean for a patient's clinical records? Will genomic information become as accessible as a person's height or weight, or will access to that data be limited to certain people? All of these questions need to be answered, or at least thoroughly discussed, before personal genomics can move into the clinical setting, McGowan adds.
If personal genomics is ever to become part of people's everyday clinical lives, continuing education is needed for both doctors and patients. "When I hang around with bioethicists, I still hear the rhetoric of fear — that we should avert our eyes because if we don't, what we learn is going to be either useless or dangerous," Duke's Angrist says. One way to combat this critique is to rethink genetics education in the US, he adds. "When you learn about [genetics] in high school, it's about some guy and his pea plants and it seems utterly remote," Angrist says. "I say that as someone who's a big Gregor Mendel fan, but also as someone who remembers suffering through high school biology and leaving high school with no interest in the life sciences because it was done in such a dry way." Instead, he suggests, genetics should be taught as something "really cool" that has "extraordinary relevance" to everyone.
In a recent editorial in the Huffington Post, Angrist wrote, "Whether [consumer genomics] initiatives lead to faster cures or diagnostic tests remains to be seen. ... The point is that they will inevitably help to lift the shroud that has enveloped human heredity in the public consciousness."
Stanford's Cho adds that clinicians are also difficult to sway when it comes to getting them to use genomics in their practices. "Until a clinician feels like something is a standard of care and there are protocols and guidelines on when to use these tests and how to use them, they won't use them," she says. "There's a pretty high bar for changing the behavior of physicians — if there's a test that clinicians are going to adopt, it's going to have to be far more useful or provide better information than what's already available."
In the end, no amount of education can change a patient's mind if they do not want to know what their genetic code tells them. "It takes a pretty strong stomach to deal with what comes out of it — it's never good news, and a lot of the information is ambiguous, and so you have to be able to deal with those kinds of ambiguities and not everybody is in a position to do that," Stanford's Quake says. "I think what you'll also find is that even if people get more educated, some people will still elect not to know. ... You have people who totally understand the potential value but they still decide that's not something they want to know about themselves, and I don't think you can say that's a wrong decision — you have to accept that."
Angrist also says that some people, no matter how well informed, will be "skittish and uncomfortable" with the idea of personal genome sequencing and will elect not to do it. "I'm the last one to say we have to force it on them," he says. "The PGP and even 23andMe and every other manifestation of personal genomics isn't for everyone in the way that bungee jumping isn't for everyone."
Despite all the problems that have yet to be worked out, and even considering the "extreme skepticism" with which some primary care doctors view genetics in general, Angrist still sees reason to be optimistic about personal genomics' chances of being folded into clinical practice. "The question five years ago was, 'Would anyone want to do this except George Church and his wacky followers?' And the answer is yes," he says. "But it's still an uphill battle."