With each new birth, complex human lives are ruthlessly pared back to their essence. Twenty-three chromosomes from a father meet with a matching set of 23 from a mother. The instructions written into the chromosomes' DNA mastermind the long, infinitely complicated process of building a brand new human being.
Nine months later, a baby comes mewling into the world. People say she's got her father's mouth or her grandmother's nose. The child grows; she's tall and green-eyed like mum. At school they say she's dyslexic, just like her aunt and several cousins.
We take it for granted that our traits, tastes, tendencies, abilities and ailments are somehow 'in the blood'.
On the flipside, we casually give genetics credit for much of what makes us unique: we note that baby siblings have different temperaments and distinct hair colours. We each receive a unique hand, dealt by our parents when they merged and reshuffled their own genetic packs.
Today, you can pay a surprisingly modest fee to have your entire genetic profile decoded and scrutinised. In eager pursuit of self-illumination, millions of consumers are buying into the young and fast-changing world of personal genomics. But there is doubt on how sharp new genetic insights will be. Inevitably, there will also be questions over the commercial value your banked DNA can have for the companies offering the tests.
We all know that cultural norms, education, family values and acts of blind chance help to shape every one of us. But still we find genetics extraordinarily compelling, even though it ultimately boils down to a code of just four symbols: the four chemical sub-units of DNA, known by their initials, A, C, G and T. Its apparent simplicity makes us believe that modern science can read and understand it.
This is why shrill fanfare greeted the unveiling of the first near-complete human genome sequence in 2000. For nearly half a century scientists had known that reading the precise order of DNA's bases was key to understanding the instructions buried within. At last, new technology allowed them to march methodically along the genome's 23 pairs of chromosomes, chalking off the sequence of As, Cs, Gs and Ts.
Our folk intuitions about inherited characteristics have lulled us into thinking genetics is simpler than it really is
Francis Collins, director of the US arm of the Human Genome Project declared: "The consequences… for our ability to understand health and disease, and a whole variety of other issues that relate to humanity, are profound."
Sequencing that first genome took most of a decade, hundreds of scientists and $3bn. A decade on, the symbols of DNA are no longer the exclusive preserve of a white-coated scientific elite. Anyone with a little spare cash can contemplate their own genetic makeup.
Californian company 23andMe leads the way in the young field of direct-to-consumer genetic testing. Under the banner 'knowledge is power', they project a utopian vision. As well as helping to trace your ancestry, 23andMe want to overhaul our entire approach to health and healthcare. They want to arm people with their own genetic risk factors and encourage them to live their lives accordingly. The grand aim is to shift medicine from a reactive to a proactive practice, with personal insight, personal empowerment and personalised care.
Your genome – the entire complement of DNA held in your 23 pairs of chromosomes – contains 3bn letters of DNA sequence. The vast majority of those letters are utterly mundane and shared, in the same place and in the same order, by every human on the planet. The remaining 1 per cent or so are more interesting. Unless you are an identical twin, the pattern of 'variants' you have inherited is a genuine one-off and gives you your individuality. It is these variants that most geneticists focus on, and that companies like 23andMe scrutinise.
I coughed up £125, spat into a plastic tube and shipped my drool to 23andMe. Their technicians fished out my DNA and used a miniature device called a microarray to examine the status of my genome. They didn't sequence my whole genome, but instead focused on some 650,000 variant positions. This process of determining whether an individual has an A, C, G or T at specific places in the genome is called genotyping.
Part of the self-illumination that personal genomics promise is the ability to reconstruct your own history. My genotype is unique, but it also marks me out as a member of a family.
The number of genetic variants I share with family members is directly proportional to how closely related we are. 23andMe help you scan for relatives who have also been genotyped. The closest they found for me were fourth cousins. The website says I share just 0.6 per cent of my DNA variants with my newfound fourth cuz Andrea del Guidice. This may be meaningful for some, but I'm not planning an emotional family reunion quite yet.
Part of the self-illumination that personal genomics promise is the ability to reconstruct your own history
Many come to genetic testing to look beyond their relatives to their deeper ancestral roots. UK company Britain's DNA has a video welcoming you to their website, featuring grim-faced Roman legionaries and bearded hunter-gatherers knapping flints and sailing in hide coracles. The dramatic voiceover asserts: "DNA takes your family history back into deep time… (and) tells the tale of the journey of your ancestors across the Earth."
But as you climb higher into your family tree, the branches become ever more numerous and fragile. In fact the whole notion of a family tree soon becomes meaningless. Chris Tyler-Smith, head of the Human Evolution team at the Wellcome Trust Sanger Institute, explains why. "You have two parents, four grandparents and so on. If you do the sums, you only have to go back a few centuries before you have millions or billions of potential ancestors. In an amazingly short time the number exceeds the number of people on the planet."
This means we all share a surprisingly small set of ancestors, surprisingly recently. According to the statistics, almost all Europeans may be descendants of the 8th-century king Charlemagne. They're equally likely to be blood relatives of the Holy Roman Emperor's lowliest serf, Viking raiders, Celtic foragers, ancient Greeks, Armenian highlanders, and every other ethnic group you can think of. It is humbling stuff.
But Tyler-Smith built his academic reputation on using genetics to reconstruct the movements of human populations in deep historical time. He explains how variants can tell where your ancestors lived tens of thousands of years ago, on a continental scale. It makes scientific sense to say your DNA is broadly 50 per cent African and 50 per cent East Asian, for example. Much of the rest is storytelling and speculation. Only if, by accident of geography or culture, your ancestors isolated themselves for centuries, could geneticists reliably tell more precise stories about where you came from.
My 23andMe report tells me my DNA is one-fifth Ashkenazi Jewish. This jars with what I know about my family. But it's estimated that today's 10 million Ashkenazim are descended from as few as 350 individuals who lived 600 to 800 years ago. This means their genetic signature is clearer than most. My father's parents were both Polish, but both claimed to be of Catholic stock. My maternal grandmother's maiden name was Franks. Jewish? This is an intriguing family mystery, but I doubt a life-changing one.
Yet it could be. The genetic signature that makes Ashkenazi heritage stand out has become entwined with some genuine health risks. Several genetic diseases crop up much more frequently in Ashkenazi groups. These include Tay-Sachs disease, Bloom syndrome, cystic fibrosis, Canavan disease and familial dysautonomia.
These are genetically determined diseases. If you inherit two copies of a single mutated gene, the disease will strike. If you inherit just one copy, the 'normal' version on the other chromosome masks the defect. Even so, you are a carrier of the disease and might pass the dangerous variant on to your offspring.
My 23andMe report lists my carrier status for 43 inherited conditions. Luckily, I'm negative for all of them. This is a straightforward relief. But if I had been a carrier, my reaction may not have been so simple. If I had cross-referenced my genetic profile with my wife's before we started building a family, how might that have affected our decisions?
The inherited conditions 23andMe report on can be serious, but are mercifully rare. As we live longer, it is heart disease, dementia, diabetes and cancer that will cripple or kill most of us. The causes of these common diseases are complex. Diet and lifestyle are important, but they all have a strong genetic component too.
If geneticists could find genuinely predictive risk factors for these conditions, then people could restructure their lives. They could regularly self-monitor for the earliest signs of disease. Medics could tailor treatments to individuals' particular body chemistry. Arguably, the knowledge that your DNA carries the risk would be the ultimate instigator for change. Anne Wojcicki, 23andMe's charismatic co-founder, responded to her genetic risk of breast cancer by reducing her alcohol consumption, for example.
What can 23andMe do to help with any of this?
If you live in America, then the answer is precisely nothing. Two years ago, US consumers could see more than 240 detailed health and trait reports, including risk factors for heart attack and type 2 diabetes. Then, at the end of 2013, the Food and Drug Administration clamped down. They implied that 23andMe's tests were inaccurate. They also worried that some consumers would misconstrue the risks, seeking out unnecessary treatments and changing their medication regimes without speaking to their doctors.
So far, regulators in Canada, the UK and several other European countries have taken a more lenient approach.
Under a section of my online report marked 'genetic risk factors' I can check my status regarding 11 carefully selected conditions. These include Alzheimer's and Parkinson's diseases, breast and ovarian cancer and some specific heart and blood disorders. Each report focuses on one or a few specific variants that are known to tilt the odds of getting sick.
Many of the risk factors sound modest, but the breast and ovarian cancer report stands out. It was finding mutations in the BRCA1 gene that led Angelina Jolie to opt for a preventative double mastectomy in 2013. Jolie's decision was brave and she inspired many women to have their genetic risk for breast cancer checked out. But some may question whether consumers should receive this sort of information without professional support. If you'd rather not know, don't send your spit to 23andMe.
To their credit, 23andMe are measured and responsible in their reporting of these risks. They are also at pains to point out that their tests are not meant to be diagnostic. Everything should be followed up with healthcare providers.
I thank my stars that nothing too sinister emerged from my profile and that there are few 'actionable' genetic risk factors in my 23andMe report. Is this the result of cautiously selective reporting by a company still smarting from FDA reprimands? Or is it a realistic reflection of the state of modern genetics?
In search of answers, I drive to the Wellcome Trust Genome Campus near Cambridge. Apart from a few spiralling box hedges near the gate, the gardeners have resisted the obvious urge to build DNA’s iconic helix into their plan. For it was here that huge tracts of the first human genome were sequenced at the turn of the century. Since then, the cost of DNA sequencing has plummeted, even as its speed has rocketed. The Trust's Genome Campus scientists have stayed at the forefront of generating and deciphering the genomic data throughout the resulting explosion of interest.
If I had cross-referenced my genetic profile with my wife's before we started building a family, how might that have affected our decisions?
Inside, Dr Jeff Barrett, head of medical genomics at the Sanger Institute, leads me through his lab – no test-tubes in sight, just rows of computer screens – and into his office. A fast-talking Bostonian, he gets straight to the point.
"We fetishise DNA and how it defines us," he says. "Genes influence everything, but our lives and our health end up being a product of our genes, our environment, personal choices and many other things."
Most of the genetic associations that 23andMe and others report on come from so-called genome-wide association studies (GWAS). For these studies, geneticists comb thousands of genomes and use statistical techniques to spot variants that crop up more often than expected in people who are taller, smarter, susceptible to a disease, or unresponsive to a certain medication.
Over the last decade, thousands of these genomic studies have been conducted. But the deluge of clear links between gene variants and medical conditions predicted by Clinton, Blair and the other human genome celebrants at the start of the 21st century hasn't materialised. Instead, we've seen a steady trickle of much more modest insights.
Last year, for example, scientists completed a large hunt for the genetic causes of schizophrenia. It involved 37,000 schizophrenics and 113,000 unaffected controls. The researchers found 108 variants which, if found in a genome, increase the probability that the owner of the genome will develop schizophrenia. This confirms beyond doubt that genetic factors contribute to schizophrenia. The rub is that the 108 variants together predict only about 7 per cent of lifetime risk of developing the condition. Each variant on its own has a tiny causal role to play.
Another huge study looked at the genetics of height. A similar take-home emerged. Height is largely under genetic control, but the genetic influence appears to be scattered confetti-like between hundreds or thousands of different variants.
In a recent essay, journalist David Dobbs described the swarm of minuscule genetic nudges that feed into complex traits like height and common disease risk as MAGOTS: Many Assorted Genes of Tiny Significance. These are the norm. Strong genetic risk factors like the mutations causing cystic fibrosis and the BRCA gene variants in breast cancer are the exceptions.
Geneticist Barrett explains why we shouldn't be too surprised by the relatively muted results of most genome-wide analyses conducted to date. Most of these studies focus on so-called 'common variants': genetic changes that tend to persist in human populations precisely because they don't cause too much upset. Over the long evolutionary timescales that have shaped our genomes, variants that caused more serious health risks made family-making tough. As a result they are much rarer.
Yet minuscule as they may be, shouldn't the risk factors caused by common variants eventually, on aggregate, be useful in helping geneticists to arrive at realistic, personalised health risk reports? Perhaps. But there is no strong reason to believe things will stack up so neatly.
Our folk intuitions about inherited characteristics have combined with mass media reporting of genes 'for' aggression, voting preference, alcoholism and the like. In fact, biological systems are fiendishly complex and built around myriad feedback loops and failsafe mechanisms. In these circumstances, the effects of genetic variants combine with and counteract each other in extremely hard-to-predict ways.
Just as crucially, many genetic effects will only reveal themselves in particular environments. Consider the Siamese cat. The toes, nose and other extremities of these elegant felines are coloured black by a pigment called melanin. But the cat’s warm underbelly tends to be pale. This is because a Siamese-specific genetic variant renders a key enzyme for melanin production heat-sensitive. Raise a Siamese in a cold room and the black pigment will spread. Turn up the radiators and you’ll have a pale puss.
Personal genomics is something of a misnomer. It's still not a very useful tool for self-enlightenment. The real power lies elsewhere
These are the uncertain probabilistic conditions that genetic testers face when they try to forecast your future health. Barrett tells me when his 23andMe report detailed a variant in the APOE gene known to increase his risk of developing Alzheimer's, he wasn't overly concerned. He knows there are many other risk factors that may or may not counteract that one.
For some, there may be a silver lining to this dense fog of genetic uncertainty. In a dystopian near future, those who seek to select and discriminate on the basis of genetic code will face the same challenges as those who want to heal. Health insurers will struggle to calculate accurate genetic risks, and therefore set realistic premiums. They still learn much more by asking how much you drink and smoke.
And as for pre-implantation screening of human embryos for the most intelligent and beautiful, it's simply not going to work.
I ask Barrett what is set to change in this fast-moving field. Will the personal risk predictions get better?
He points to another source of genetic variation, more elusive than the common variant. All of our genomes contain a smattering of much rarer changes. Some of these genetic anomalies will have more disruptive biological effects, and thus much more predictive value. But these rare variants are too hard to find and understand using just the genotyping microarrays employed by 23andMe and their competitors. To find the rare stuff, the whole genome needs to be sequenced, letter-by-letter.
Sequencing costs are plunging towards the $1000 mark some deem 'affordable'. At the moment, 23andMe don't offer whole genome sequencing but, on the phone from California, their development director Emily Drabant Conley hinted that this service will come online soon.
Jeff Barrett cautions against our assuming this will be a step forward: "It seems somehow 'sexier' to have whole genome data. Some people thought we'd understand it all very quickly. But actually the reverse is true. You need bigger datasets, containing even more genomes, to tease out what the rare stuff does."
Driving back home from the Genome Campus it dawns on me that, as things stand, 'personal genomics' is a misnomer. It's still not a very useful tool for self-enlightenment. The real power lies elsewhere: in the database.
With over a million customers profiled worldwide, 23andMe have the biggest, most coherent genetic database in the world. They have also become adept at amassing the extra layers of information that dramatically enhance the predictive value of your banked DNA.
When you send your spit, they ask if you want to participate in genetic research. Curious and eager to contribute to science, over 80 per cent say yes. Having done so, they fill in surveys and report on their family and medical histories; and then answer a bizarre array of questions from "Do you cry easily?" to "How often do you jiggle your leg?"
23andMe feed all this data into their vast, burgeoning databanks. The largest genome-wide meta-analyses conducted to date involved a quarter of a million genomes. 23andMe have nearly four times as many in hand. This is their key asset.
Even if genetic variants serve as weak predictors of an individual's health risk or psychological profile, their potential medical and commercial value is huge, especially in drug development. In some cases, new drugs will target patients with specific genetic profiles, so-called precision medicines. But often the genetics will be the clues that get the big pharma machine rolling towards more widely applicable, and widely marketable, therapies.
A current example involves FTO, a gene involved in fat processing. A variant of this gene is somewhat predictive of obesity. Having found the variant, researchers set out to establish what the gene actually does. Now a team at Harvard think they have an excellent target for anti-obesity drugs, that could work whether people carry the FTO variant or not.
23andMe have never tried to conceal that drug development has always been a big part of their game plan: they recently announced an in-house therapeutics division and have also agreed a series of high-profile partnerships with pharmaceutical players. These companies are paying millions of dollars to peek into 23andMe's unique database. With luck, new drugs and therapies will emerge from these efforts.
For some, 23andMe's approach to business is problematic. Donna Dickenson is emeritus professor of medical ethics at the University of London. She likens their approach to that of Tom Sawyer, who convinced his friends to pay him for the privilege of painting his picket fence. "There's nothing wrong with people being committed to advancing science, that's a great thing. But here people are paying to be volunteers," she says.
"If [23andMe's database] is to be a communal resource, maybe we need some communal protection," she suggests. Dickenson thinks 23andMe's customers should have more say in who can see their information and who can profit from it.
Dickenson points to the parallels with internet corporations like Google and Facebook. In each case, consumers trade intimate data for access to information and services. In 2013, journalist Charles Seife raised the possibility that this barter of genetic information could have some unanticipated results. Seife wondered whether part of 23andMe's business, like that of the online giants, may soon come from advertising.
While many of us accept, some even welcome, targeted adverts that offer us products closely aligned with our interests and social lives, targeted health-related advertising seems different. Unrequested reminders of genetic risk factors could induce worry and hypochondria. In a market-driven health sector, this could lead to unnecessary treatment.
23andMe's Emily Drabant Conley doesn't rule out the possibility of using of genetics in advertising, "but it only makes sense to do that in a model where the consumer has opted in and that's what they want." She insists that 23andMe never shares identifiable, individual-level data with third party organisations without explicit customer consent. She also confirms that the company has no current plans to use this information for advertising. Genetically targeted adverts don't seem to be on the immediate horizon.
In reality, we still know little about how people respond to learning about their genetic risk factors. In University College London's Cancer Institute, I meet someone who wants to find out.
Stephan Beck, professor of medical genomics, heads the UK offshoot of the Personal Genome Project (PGP). Set up in Harvard in 2005 by influential genomics pioneer George Church, the PGP vision is to create an open-access database containing hundreds of thousands of genome sequences. The project will match each genome with medical records and self-reported health and lifestyle information.
PGP participants, who are all volunteers, are sometimes described as health information altruists. They take a bold stance on genetic privacy: let it all hang out. All the data they contribute to the project is freely visible and downloadable online. It's all anonymised, but in practice it wouldn't be hard to track down people's real identities.
Beck explains the unique function of the growing international network of PGPs. "We're open-access, but we're not hackers," he says. "We're recruiting large cohorts of well-informed and highly engaged volunteers. This lets us test the boundaries of what [genomic information] should be reported and lets us find out what people experience as detrimental."
Genomics can revolutionise healthcare, Beck believes. But to realise its potential, the data must be liberated. He has no specific problem with the commercial model. "23andMe is becoming an important database: it's very difficult to get a million cohort and their worldwide sampling is fantastic," he says. But he passionately believes that all public sector research should work for the public good.
Health insurers will struggle to calculate accurate genetic risks, and therefore set realistic premiums. They still learn much more by asking how much you drink and smoke
Beck is pained by the vast amount of DNA sequence data currently languishing in private hard-drives. "Probably 95 or even 99 per cent of all clinical sequencing will never see the [open-access] light of day. We spend billions to generate this data and then we lock it away in the bottom drawer. It's insane."
A key obstacle is the current informed-consent model that most ethics committees impose on research projects. When subjects agree to have their DNA sequenced, the scientists must specify exactly how they will use the information. Typically, they do the study and never look at the data again. Byzantine layers of paperwork obstruct subsequent researchers from analysing the data.
Beck is convinced that this is not what people want. Within a week of launching PGP-UK in 2013, 10,000 volunteers signed up, overwhelming their enrolment system. The 800,000-plus opting in for 23andMe research further demonstrates that people are curious about genetics and want to help advance research. Once sequenced, DNA data can be used again and again.
Looking ahead, both Beck and ethicist Dickenson think a more dynamic form of consent may serve everyone's interests best. Under this new framework, researchers and participants would be in constant dialogue and data volunteers could easily track and manage their genetic and health information.
Since human genomics projects must crowd-source their data, the whole endeavour is built on trust. A system that puts transparency and personal control to the fore in this way has to be good. It should apply to both public and corporate sector genomic research.
Personal genomics may never provide compelling answers to our soul-searching "Who am I?" questions, but the science will grow in power. It will help design new drugs and guide an increasing number of reproductive and healthcare decisions. UK secretary of state for health Jeremy Hunt has even voiced his enthusiasm for sequencing the genomes of all British babies at birth.
Faced with a future where genomic profiling is commonplace, or even mandatory, one might ask, "What can we actually do with all these genetics risks? We can't change our genes, so does it really help to know what they say?"
But according to some forecasters our genetic inheritance may no longer be set in stone. Biologists have recently developed a new set of surgically precise tools for editing DNA sequences. The torchbearers for new technologies like CRISPR claim that DNA editing could 'provide unlimited clean energy' and 'solve world hunger'. The strapline of a recent feature in Wired magazine promised it could 'eliminate disease'.
Most of this sort of hype ignores the huge technical and ethical obstacles that need to be overcome before human DNA can be edited. But assuming these can eventually be cleared, Jeff Barrett is hopeful that the likes of Huntington's disease, cystic fibrosis and haemophilia could be deleted from DNA sequences. "If you could cure all those, then that would be a huge change in human history," he says. "That isn't on the immediate horizon but it's not so far off."
This is a genuinely exciting possibility. Yet the reality is that these high-impact disease-causing variants blight the lives of a fairly small slice of humanity. Look beyond and DNA editing seems less of a solution.
Liberating for some, frustrating for others; a key message stands out from genome science. The genetic influences that shape our complex human lives in health and in disease are, in turn, fantastically complex.
Even with the keenest DNA editing tools imaginable, the combined intricacy and scale of the problem would overwhelm most attempts to fix or embellish. It'd be like trying to tackle a seething infestation of maggots by picking them out, one at a time, with a needle-sharp pair of surgical forceps.