Ever since DNA was first used in 1986 to catch a killer, it's swashbuckled its way through society as an almost infallible weapon of truth: convicting the guilty, freeing the innocent, revealing bloodlines, paternity and identity.
On the health front, even before it had a name, excitement over DNA's power straddled centuries. But only the 21st has cast it as the attainable key to "precision medicine" – enabling doctors and patients to practise pro-active care: pinpointing diseases before they strike, and fighting them with targeted therapies tailored to an individual's unique genome.
Yet, the more genomes researchers study, the more evidence mounts that using DNA to predict health risks is anything but precise – for now, at least. These days, nature's longest thread comes off a like a trickster, shape-shifting from person to person, written in a wily language no one fully understands, with at least three billion ways to misread it.
To try to bridge the chasm between reading and comprehension, scientists leading Canada's Personal Genome Project say they have taken the deepest dive possible into human DNA, conducting the most thorough analysis that current computing allows on the whole genome sequences of 56 Canadians who agreed to take part in the unusual and continuing experiment. Their report is published today in the Canadian Medical Association Journal.
Launched publicly in Canada in 2012, the PGP aims to build an open, online database of Canadian genomes for use by researchers anywhere and more than 1,100 Canadians have signed up so far. But by unravelling the entire codes of just the inaugural participants, what the project leaders have found is both promising and perplexing: medically relevant information in each volunteer, but also a vast trove of mysterious quirks and dramatic glitches that none of them expected to see in a cohort of healthy people.
Surprising, too, is that many of the volunteers carry mutated genes that suggest they should be sick, diseased or even near death – but aren't.
Take Participant No. 16, whose genes seem to tell the grim story of an aortic stenosis. The potentially lethal heart defect develops before birth, narrowing the body's main artery just above the valve that connects it to the heart, forcing it to work harder, which can lead to chest pains, heart failure and the risk of sudden death. On paper, No. 16 looks like a time bomb. In real life, he is a healthy 67-year-old who works long hours, skis, trains with CrossFit three times a week and, according to a CT scan, has a normal heart.
Participant No. 27 is another jolt. Missing from about 70 per cent of her blood cells is an entire chromosome, the X sex chromosome, of which women usually have two copies. On paper, No. 27 has mosaic Turner syndrome, a rare disorder that only affects females and can lead to impaired sexual development, short stature, a webbed neck and heart problems. In reality, she is a healthy 54-year-old who had no inkling her DNA harboured such a twist.
The report, compiled by 53 researchers at the Hospital for Sick Children and the University of Toronto, captures the staggering breadth of how much is left to learn before science can properly interpret humanity's operating code. It offers proof that most research to date has overlooked the different types of genetic glitches that can lead to health problems, while adding to the growing realization that many genes once thought to cause disease probably don't.
All of which implies that in these foggy days of the so-called genomic revolution, people can easily be led astray by test results that may bring false comfort, or needless stress and treatment – especially now, when the largely unregulated market of direct-to-consumer DNA tests is booming.
Just ask Participant No. 16 about the heart abnormalities his genes suggest.
"You can imagine someone getting this report back and saying, 'Oh my God, I'm gonna die of this! How do I check it out?' Then there will be echocardiograms and all these things being done on people who are perfectly healthy," says Doug Mowbray, who wasn't troubled by his vexing results because the participants had a genetic counsellor walk them through the fog. But also because he's a doctor himself – a radiologist, in fact, in Southwestern Ontario. He knew his heart was normal because he took a good look at it while calibrating a new imaging scanner a few years ago.
Unsurprisingly, the new report concludes that Whole Genome Sequencing (WGS) is not ready to become a standard feature of routine health care. But the researchers predict that "despite a considerable burden of uncertainty, and the possibility that false-positive findings may engender follow-up investigations and a 'worried well' population" – it will be: the plummeting cost of reading DNA is pushing it into mainstream medicine.
Study leader Stephen Scherer, director of both the McLaughlin Centre at U of T and the Centre for Applied Genomics at Sick Kids, says people send him e-mails almost daily asking if they should have their whole genomes sequenced. His answer: "At the current time, unless you have a medical query that can't be explained by traditional means … don't. But when the price comes down, into the hundreds, or less, there's not going to be a question any more – it's just going to happen. There's going to be an onslaught of information, and society needs to figure out how to deal with that."
Cutting costs, saving lives
In less than two decades, the cost to sequence a human genome has dropped by billions of dollars. The first map of the genome, decoded in 2000, rang in at about $3-billion. Today, such a map costs roughly $1,000. Estimates are that in five years, both researchers and the curious will be able to see what they're made of for $100 – less than the cost of an X-ray.
At hospitals and medical labs across the country, the sequencing of whole genomes, or parts of them, is taking off, just as it is in the consumer market. While most of it remains in the research sphere, doctors are increasingly looking to whole genomes to diagnose disease, or to choose the best drugs for their patients – and it's making a difference: unearthing the roots of rare conditions for patients who have waited years for answers; pointing the way to effective medicines; and predicting if patients will tolerate them. In cancer care, especially, it's transforming treatments and early diagnosis.
So, filling the knowledge gap is urgent not only to counter the potential for misleading results, Dr. Scherer says, but because the chance increases all the time that you will find something helpful in your genome, something that might even save your life or the lives of your children.
In all, the study found that a quarter of the participants, or 14 people, carry genetic variants associated with heart disease, cancer and neurological conditions. Nearly a quarter were also found to be at risk of potentially life-threatening reactions to certain drugs. More than 94 per cent, or 53 of the 56 volunteers, were carriers of rare disease genes – such as cystic fibrosis – that give a one-in-four chance of having a child with that disease should a co-parent also be a carrier.
The researchers recommended further monitoring, tests and counselling for several participants, and noted the findings have health-management implications for their families and future generations. As. Dr. Scherer explains, decoding a whole genome may happen once, but individuals, and their descendants, may turn to it many times for answers and explanations.
Genetic testing's quantum leap
A vast undertaking
From the get-go, the Personal Genome Project was unorthodox – in part because participants were not only asked to share their genomes online but also personal details about their lifestyles, traits and medical histories. Since DNA is a living language that responds to how we live, and even where we live – studies find it can change with the weather – the project has given researchers the capacity to break down the lifelong bossa nova that genes dance with the environment.
The philosophy behind the project sprouted at Harvard University, where the first PGP project began in 2005. But when Dr. Scherer started making plans in 2007 to launch a Canadian version, it was unclear how many Canadians would go in for that kind of molecular Full Monty, especially with no guarantees of privacy. At the time, Canada was the only Group of Eight country without a law to protect genetic information from use by a third party. Before signing on, participants had to pass an exam to demonstrate they understood the wide range of risks involved – that they could lose the ability to obtain life insurance or a job, discover they're not biologically related to their families or that someone could even make a synthetic copy of their DNA and plant it at a crime scene.
After the Canadian launch of PGP, the movement gathered global steam, with projects taking off in Britain, Austria and, last October, in China. But in Canada, with no dedicated funding at the start, the first 25 participants had to pay for themselves to be sequenced – which for most, in 2012, cost about $4,000. "These volunteers were the heroes," Dr. Scherer says.
But asking people to pay their own way also resulted in a fairly select group of early volunteers. Most are affluent, well educated and, the paper notes, "idealistic about the promise of genomic medicine" – and, it so happens, white. Apart from one participant from East Asia, one described as "white native" and three from the Middle East, all are Caucasian. The 25 women and 31 men ranged in age from 25 to 81 at enrolment, the median age being 51. The researchers are up front about the study's lack of diversity and say in the paper that they "explicitly aim to expand" it in the next rounds of analysis, with participation from Indigenous people and recent immigrants.
The initial selection process was also skewed deliberately, Dr. Scherer says, toward enrolling people "who are in a position to spread the word about the experience." Many of the participants also work in health care: doctors, professors, genetic counsellors; even more unusually, some are also co-authors on the paper.
Progress was slow going in the early years, but in 2016, after a new generation of sequencing machines arrived at Dr. Scherer's lab, and the McLaughlin Centre put up funding, the team decided to reread the genomes of the first participants at a level never before possible. Then, last March, Dr. Scherer decided to freeze the data at 56 samples, "knowing it would take about four good months to analyze."
He wasn't wrong. Trying to make sense of the various quirks that turned up in about a hundred billion units of code from the sequences of those 56 demanded a huge research team that included specialists in cardiology, biology pharmacology, epigenetics, paediatrics, metabolic disorders, bioinformatics and genetic counselling.
Sick Kids scientist Miriam Reuter, lead author on the paper, says it was not clear how to approach all the DNA sequences they had generated, but they decided to home in on everything that looked as if it could affect health – the kind of patient information doctors may face in the near future. Diane Kelsall, editor of the Canadian Medical Association Journal and an Ottawa family doctor, describes the results as a vital resource for physicians already fielding questions from patients who have either taken commercial tests or are thinking about it. "This will enable us to counsel them," Dr. Kelsall says. "This is a groundbreaking project for Canada."
How your DNA works
HUMAN CELL
Nucleus of every cell contains 46 chromosomes, 23 from each parent
CHROMOSOME
They are comprised of one tightly coiled molecule of deoxyribonucleic acid (DNA) with proteins that serve to package DNA and control its functions
Nucleus
DNA
Carries unique genetic code that determines characteristics of each person. It’s made of chemical bases A, C, G and T. Each A base bonds with T base and each G base with C
GENETIC CODE
Order of nucleotides within gene is inheritable instructions needed to make protein molecules
PROTEIN
Some proteins are building materials of cells – skin, heart, blood – while others control biological processes such as digesting food or carrying oxygen in blood
VARIANTS
Variant refers to any change in sequence and is not necessarily harmful. Only when the change results in damaging the proteins a gene makes are they considered to be the possible cause of a genetic disease. Variants are described on a spectrum as being "likely benign, benign, likely pathogenic
or pathogenic.”
THE GLOBE AND MAIL, SOURCE: GRAPHICNEWS
HUMAN CELL
Nucleus of every cell contains 46 chromosomes, 23 from each parent
CHROMOSOME
They are comprised of one tightly coiled molecule of deoxyribonucleic acid (DNA) with proteins that serve to package DNA and control its functions
Nucleus
DNA
Carries unique genetic code that determines characteristics of each person. It’s made of chemical bases A, C, G and T. Each A base bonds with T base and each G base with C
GENETIC CODE
Order of nucleotides within gene is inheritable instructions needed to make protein molecules
PROTEIN
Some proteins are building materials of cells – skin, heart, blood – while others control biological processes such as digesting food or carrying oxygen in blood
VARIANTS
Variant refers to any change in sequence and is not necessarily harmful. Only when the change results in damaging the proteins a gene makes are they considered to be the possible cause of a genetic disease. Variants are described on a spectrum as being "likely benign, benign, likely pathogenic or pathogenic.”
THE GLOBE AND MAIL, SOURCE: GRAPHICNEWS
HUMAN CELL
Nucleus of every cell contains 46 chromosomes, 23 from each parent
CHROMOSOME
They are comprised of one tightly coiled molecule of deoxyribonucleic acid (DNA) with proteins that serve to package DNA and control its functions
Nucleus
DNA
Carries unique genetic code that determines characteristics of each person. It’s made of chemical bases A, C, G and T. Each A base bonds with T base and each G base with C
GENETIC CODE
Order of nucleotides within gene is inheritable instructions needed to make protein molecules
PROTEIN
Some proteins are building materials of cells – skin, heart, blood – while others control biological processes such as digesting food or carrying oxygen in blood
VARIANTS
Variant refers to any change in sequence and is not necessarily harmful. Only when the change results in damaging the proteins a gene makes are they considered to be the possible cause of a genetic disease. Variants are described on a spectrum as being "likely benign, benign, likely pathogenic or
pathogenic.”
THE GLOBE AND MAIL, SOURCE: GRAPHICNEWS
We're all mutants
The PGP's deep dive into human DNA overturns the notion that the healthy inherit from their parents two matching sets of 23 tidy chromosomes that line up like soldiers. Indeed, it finds major structural differences not only between people, but between chromosomes in the same person. "People are walking around with these massive chunks of chromosomes missing, and [duplications] that affect multiple genes," Dr. Scherer says. "When I was a student, these things were associated with Down Syndrome, or DiGeorge Syndrome, so to see how structurally different we are, and how many examples there are in this [small, healthy] group – it amazes me."
Most research on DNA has focused on the single letter "typos" in genetic code as signposts of disease. But the PGP study concludes that further research should include places where whole "sentences" or "paragraphs" of code may be missing, or upside down, or, like a stutter, repeated – a quirk known as a copy-number variation, or a CNV.
The PGP study found more than 27,000 CNVs among the 56 Canadians. The tally includes eight participants missing huge swaths of code from their genomes; a man with a long stretch of Chromosome 20 that's inverted; and, in Participant No. 56, such a dramatic hiccup in Chromosome 17 that she has extra copies of a full 16 genes. "It could make her special, it might eventually make her sick," Dr. Scherer says. "It's definitely doing something; we just don't know what it is. There's nothing to compare these copy-number variants against."
Yet, without that information, there's a serious risk of missing a health issue: "It would be like going to your doctor's office to get checked out for skin cancer, and they're looking for moles of a certain size – but miss the great big gaping tumour on your back."
The study also examined changes in mitochondrial DNA, which people inherit only from their mothers. In one participant, this separate ring of genetic code, which cells use to convert food to energy, contained an abnormality researchers had previously predicted could lead to neurological disorders in childhood. But there was no evidence of it doing so.
Dennis Bulman, a molecular geneticist at the Children's Hospital of Eastern Ontario, who was not involved in the research, said the depth of the study, and the fact that it was done on "normal" volunteers, was long overdue. "It's a comprehensive piece of work," said Dr. Bulman, who's also a senior scientist with the Ontario Newborn Screening Program. "It's about time something like this was done. They've annotated everything."
While the study is small, and it can't be extrapolated to the general population, Dr. Bulman noted that it took nearly as many researchers as there were participants to sift through the data. It may come down to "the $10 genome, and the $100,000 analysis," he quipped.
A team of researchers is hoping to create a system for First Nations people to gather and oversee their own genetic data in order to improve diagnoses and health outcomes related to genetic disease, Ivan Semeniuk reports.
Biases – and fallout
The PGP identified variants in the study by comparing them with the third version of the very first human-genome map, which was compiled from the chromosomes of 13 people in Buffalo. Then, researchers checked to see whether those variants had been tagged as potentially harmful in various disease-gene databases. But there's a wide reckoning in the field that much of this database information is lopsided: Ever since the first disease gene was identified in Canada in 1989, researchers have generally hunted for disease mutations by looking at people who have a particular disorder, or affected family members.
That made for biased samples, says Heidi Rehm, an associate professor of pathology at Brigham and Women's Hospital and Harvard Medical School. "The significant challenge," Dr. Rehm says, "is to have a true estimate of the likelihood of a healthy person developing disease if they carry the variant, a true estimate of penetrance."
Ultimately, it comes down to a judgment call as to whether there's enough evidence in the disease-gene databases to flag a variant as dangerous – and it's a wide , wishy-washy spectrum of judgment that ranges from "likely benign" to "pathogenic" or, as is often the case, "a variant of unknown significance," which basically parks it in diagnostic limbo.
Dr. Rehm estimates there are more than 100 million variants that have been identified in public databases, while only about 330,000 of them have been interpreted, and "probably two-thirds of those are of uncertain significance."
In the United States, proof that the knowledge gap can lead to serious misinterpretations has already surfaced. In South Carolina, a wrongful-death lawsuit is under way against a private testing company for allegedly misclassifying a mutation that may have cost a two-year-old boy his life. The boy's mother, Amy Williams, claims that the company misinterpreted a mutation in her son, Christian, who suffered from epilepsy. As a result, Ms. Williams says he did not receive the effective treatment he could have, and was instead given medication that made him worse. Christian died following a severe seizure in 2008.
In 2016, doctors at the Mayo Clinic detailed the troubling case of two dozen family members who were wrongly informed by a testing company that they carried a heart-related mutation that could cause sudden death.
In a video release from the Mayo, genetic cardiologist Michael Ackerman notes that "the fumbles that are occurring with genetic testing out there … are starting to mount."
As are concerns around a consumer marketplace that waits for no one. Mail-order DNA tests, from companies such as 23 andMe and Ancestry.ca, have helped fuel a market that's tripled in size in the past four years. These firms only parse the genome for certain markers, to tell their customers if they're lactose intolerant, say, or hail from Viking stock. But over the next several years, as the cost of sequencing whole genomes falls to practically nothing, forecasts suggest that genetic testing will become a $10-billion global industry.
A mix of signals
In the genomic era, people keen to dip into their DNA should be comfortable curling up with ambiguity. Chances are, as the PGP study participants found, you may well carry genes for conditions you don't have, while discovering that there's no explanations for those you do.
Logan Donaldson, a 49-year-old York University biology professor, had no idea when he joined the PGP that he would go on to develop early onset glaucoma in 2015. The condition requires him to take daily medication and to have his eye pressure maintained at normal levels by laser therapy. Some might chalk it up to bad luck, but, Prof. Donaldson says, "since I had my complete genome, I started scouring it for some explanation."
He discovered he carries a rare and serious mutation in one of the seven genes linked to glaucoma. Still, "as exciting as that might be," he says, there's not enough data to be confident it's the cause without further testing and investigation.
Nor can his genome currently offer him an explanation for the low cholesterol that runs in his family.
Prof. Donaldson's levels are two times lower than normal, which has no apparent effect on his life, beyond the jokes he gets to share about undergoing "intensive french-fry therapy."
Prof. Donaldson, who has far more genetic literacy than most, says his genome has left him with more questions than answers, but he feels that most people , "even if you only had high-school biology going into this," are not likely to overreact to what they find in their genetic closets: "I think people are smarter than that."
The American College of Medical Genetics advises that people should be informed of health risks uncovered in their genome only if such risks are medically actionable, which currently includes a fluctuating list of more than 50 conditions. By contrast, the very aim of the PGP study was to dig up every relevant health issue that could be gleaned from a genome and to gauge how participants react to the mountain of murky data.
But, given the profiles of the early recruits, few had high expectations for themselves going into it. "I did this to contribute something for the long term, to help build a big database of thousands of people," says Howard Gaskin, a Toronto pharmaceutical executive now in his late 50s. In 2015, Mr. Gaskin's heart stopped after running a half-marathon in Brooklyn, N.Y. He collapsed at the finish line. Yet nothing so far in Mr. Gaskin's genome predicted the event.
To Dr. Mowbray, Participant No. 16, the real danger in the evolving field is among those who dive solo into their whole genomes: "If people are going to do this genetic decoding but don't have the proper genetic-counselling follow-up, they're going to be stirring up all kinds of dismay."
Dr. Mowbray was found to carry a total of "26 genetic defects," he says, none of which he has, or at least, not yet. Beyond the gene linked to aortic stenosis, he also harbours a variant tied to premature ovarian failure, but, having no ovaries, he "automatically discounted it." The analysis also showed he has a dominant gene variant for developing schwannomas, benign yet often painful tumours that form on the sheaths around nerves. But he's never had one, nor does he have a family history of the disorder.
Dr. Mowbray sees his string of false positives as evidence that the research is important and necessary, which is why, as a man close to retirement, without any children, he joined the project. "They know that these defects cause genetic illness, because that's how it was first identified – in people who had the illness. But what they don't know is how many people have these defects that don't have the illness – and the question is why," he says. "Maybe the thing that's stopping it in me may point them in a direction for treating people who may have this in the future."
His results do have some utility. Dr. Mowbray learned about a medication he should avoid, and that he carries the variant for an irregular heartbeat disorder both his mother and grandmother had. He suspected he would eventually develop it, but since learning of his genetic risk, he takes a daily Aspirin.
Even Michael Szego, the project's lead ethicist, discovered something both leading and misleading in his genome. He's a carrier for Tay-Sachs disease, which may be relevant to his children when they reach child-bearing age. But his genes also suggest that they're children he shouldn't have: His DNA carries a variant tied to hypogonadism, a condition that impairs masculine development and leads to infertility. Dr. Szego is a healthy father of three.
There may be various reasons the participants appear to be healthy despite the pathogenic variants they carry, lead author Dr. Reuter says. It could be that symptoms of a condition don't strike until later in life or, as Dr. Mowbray suggests, something in his genome is protective. Or, Dr. Reuter adds, the variant's link to disease has been overestimated, or simply wrong.
Dr. Scherer stands among the genomic sequencing machines in the TCAG lab. Dr. Scherer is skeptical of whether humans will ever fully understand their own genetic code, and suggests the major breakthroughs will have to wait until artificial intelligence is advanced enough to decipher the genome.
A job too big for humans
Last summer, in a speech to science graduates at the University of Waterloo, Dr. Scherer described the high-speed progress he's seen in genetics in the past 30 years: how, as a student, it took him eight months to sequence one gene in a roundworm, a job that would now take eight hours. He told them the 10-year hunt for the cystic-fibrosis gene at Sick Kids would now take 10 weeks. "The most successful scientists of my generation were the ones who built the machines to generate massive novel data sets," he said, allowing researchers to read the language of DNA that has evolved to create human minds "capable of decoding our own instruction manual."
Yet, Dr. Scherer suspects that humans will never be able to truly translate that language. He predicts it will be the advent of artificial intelligence that will yield the next big genomic breakthroughs; in the meantime, humanity will have to rely on its machines to help it connect the dots.
The world seems determined to do so. The United States is building a one-million-genome-strong database under its Precision Medicine Initiative. With its brand new Big Data Nanjing Centre, China will be capable of decoding half a million genomes a year. In December, Finland announced FinnGen – a six-year project to combine the genomes and digital health records of 500,000 Finns. Britain is at work on the 100,000 Genomes Project; Turkey has launched its own project, and Israel's is in the works.
Canada has no other open-source effort to sequence its citizens beyond the Personal Genome Project, which remains a small, sideline operation for most of the researchers involved. But, Dr. Scherer says, the country should have something like "PGP on steroids."
Still, while big national genome projects are necessary to propel the basic science, the market will outpace these efforts, says Andrew Hessel, PGP Participant No. 10, a biotech entrepreneur who now lives near San Francisco. A self-described futurist, Mr. Hessel predicts that human genomes – and lots of them – will accumulate naturally in the commercial sphere, as sequencing costs keep dropping and more people buy into the information their DNA contains.
"Who has been the most successful in getting people to open up their DNA? It's been the companies, Ancestry.com, 23 andMe, who have millions of samples in their databases," Mr. Hessel says. "Companies will find a way to mine that data, whether it's genomic research, drug research or consumer genomics. … Procter & Gamble may use it to sell you toothpaste. The Mayo may use it for research."
The idea is not far-fetched. 23 andMe is selling access to the DNA in its customer database (which customers themselves pay to be part of) to several drug companies. One, Genentech, paid a reported US$10-million for a look at the genes of people with Parkinson's disease.
"I fully expect that, in the near future, sequencing genomes will fall to free," Mr. Hessel says, "and some people's genomes will even be worth paying for because they have some interesting feature, trait or flaw" – a rare disease, or are high achievers, or simply because they're celebrities. "We're going to see a completely new dynamic appear that could potentially bring in millions, tens of millions, potentially hundreds of millions of genomes over the next decade."
"It is a revolution," he says, "it just takes a while for a revolution to build up steam and change everyone's lives."
THE GENETIC FUTURE: MORE FROM THE GLOBE AND MAIL
