n a recent trip to Boston, I had breakfast with a scientist named Manolis Kellis. He met me in the little dining room at my hotel, and we grabbed plates to load with food. Immediately, Kellis was enchanted by the waffle iron. He started pouring a waterfall of batter onto the hot metal and piled up a mound of waffles on his plate.
“You have to help me eat these,” Kellis announced when we sat down. He spoke with a vestige of an accent; he was born in Greece as Manolis Kamvysselis, and came to the United States in 1995 to go to MIT for college. He never left; today he’s a waffle-loving professor of computer science there.
Waffles — or rather the calories they contain — were very much on my mind. I was meeting with Kellis so he could explain how one of my genetic variants puts me at risk of getting fat — a variant that Kellis carries, too.
As I wrote the other day, I recently had my genome sequenced and got my hands on the raw data. Since then, I’ve been enlisting scientists to help me understand the secrets it contains.
I began by looking at the genetic variants that scientists understand best. These are the mutations that alter a single base in a protein-coding gene. But these variants are limited to a tiny portion of my genome. Only about 1 percent of human DNA is made up of protein-coding genes.
I was talking to Kellis about the kind of variant that lurks in the other 99 percent: the mysterious wilderness known as non-coding DNA.
Non-coding DNA has puzzled scientists for decades. They’ve struggled to determine which parts affect our health, and which do nothing at all. That uncertainty has led a lot of scientists to focus just on protein-coding genes, because that’s where they have the best odds of discovering something that can help people.
But Kellis and others want to find the thousands of segments nestled in non-coding DNA that have important functions.
In recent years, Kellis and other researchers have developed methods to distinguish possible jewels of non-coding DNA from all the surrounding junk. And they have gotten so good at identifying these possible elements in non-coding DNA that they’re now looking for genetic puzzles to solve.
Recently, Kellis decided to look at one of the biggest of them all: the link between obesity and a gene called FTO.
“It’s just screaming, ‘Please study me!’” Kellis said.
In 2007, two teams of scientists independently discovered that an unusually large number of obese people shared variants in the FTO gene. It was the first time scientists had linked a gene to obesity, and subsequent studies only strengthened the connection. The variant that Kellis and I both carry, known as rs1421085, causes people to put on an average of 7 pounds.
After this discovery, a lot of researchers started trying to figure out how exactly FTO influenced body weight. They observed that brain cells make a lot of FTO proteins. Some researchers hypothesized that the proteins might control hunger. But researchers couldn’t find any firm evidence for that idea.
Further research revealed that rs1421085 was not actually in the FTO gene. Instead, it was located in nearby non-coding DNA. That discovery raised the possibility that rs1421085 altered people’s weight in a way that didn’t involve FTO at all.
Kellis collaborated with Melina Claussnitzer, a visiting professor from the University of Munich, to look at the non-coding DNA in which rs1421085 was located. They found that it’s actually located in a genetic switch that had previously been unknown. Even more intriguingly, they found that the switch activates several different genes in fat cells all at once.
Having discovered this switch, Kellis dubbed it Obesity1. “I call it OBE1 for short, in honor of Obi-Wan Kenobi,” Kellis said. “You know, Obi-Wan Kenobi was a little chubby as well.”
Kellis and his colleagues then investigated human fat cells to see how OBE1 influences people’s weight. The genes it switches on cause fat cells to store fat away in reserves. If OBE1 turns off, the cells burn the fat instead.
“It’s like a switch on the wall, and for the people with the risk variant, it’s stuck in the on position,” said Kellis. “But for the non-risk people, you can switch it on and off.”
Kellis and his colleagues suspected that the genes controlled by the OBE1 switch had a powerful influence over body weight. To test that idea, they engineered mice to shut down one of those genes. The engineered mice lost half their body fat and no longer put on extra weight on a high-fat diet. Kellis and h is colleagues found that they burned more energy in their fat cells, even in their sleep.
Kellis and his colleagues then manipulated the mice, interfering with OBE1. Now his mice could produce heat from their fat cells like normal animals. Kellis and his colleagues found that if the mice ate a high-fat diet, they stopped putting on extra weight.
Discovering this series of steps in our metabolism could point to new drugs to treat obesity, according to Kellis. “You suddenly have levers and handles you can pull,” he said.
When Kellis discovered he had the fat-boosting variant of OBE1 a few years ago, it made a certain amount of sense to him. “At some point I was chubbier, and I had to work hard to lose some weight,” he said. And now his new research is giving him a deeper understanding of his biology.
Kellis feels that watching his diet is effective enough to cope with his own version of OBE1. “There’s no downside,” he said. But dieting is not enough for some people. Now that we know OBE1 may be important to how we store fat, Kellis said, it may be possible to manipulate it with drugs. Instead of adjusting people’s weight by a few pounds, as natural genetic variations do, a drug acting on OBE1 might be able to produce more dramatic weight loss.
“We may be able to help people who are losing the battle,” Kellis said.
I may not be obese, but I wouldn’t mind losing a few pounds. If I’m not losing the battle against OBE1, the best I can say is that I’m in a stalemate with that powerful bit of non-coding DNA. So I think the next time Kellis offers me some waffles, I’ll turn him down.
fter the science of genetics was born in the early 1900s, the first generation of geneticists had to work with both hands tied behind their back. They didn’t know that genes are encoded in DNA, or how changes to its sequence could change the traits that people inherited. As for technology, they didn’t have much beyond crude microscopes and simple chemical stains.
And yet, despite their handicap, they made a string of fundamental discoveries about genetics.
One of those discoveries was that DNA isn’t perfect. When cells copy themselves, passing along traits to offspring, they sometimes make mistakes. Sometimes, they make small mistakes. But sometimes they screw up in a big way. Cells may accidentally add extra bases to a new DNA molecule. They may even duplicate thousands — sometimes even millions — of bases. If the genome is a bit like a cookbook for making an organism, these duplications are like printing extra pages by mistake.
They can also screw up in the opposite way, skipping over a long stretch of DNA, producing a new genome without it. In these cases, they are neatly slicing out pages from the genomic cookbook.
These big duplications and deletions are collectively known as copy number variants, and they can have profound consequences for our health. In 1959, for example, scientists discovered that Down syndrome is caused when the full sequence of chromosome 21 gets duplicated.
Unfortunately, even though scientists have known about these variants for 80 years, scientists today don’t understand them very well. The most common techniques for sequencing DNA — chopping it into tiny fragments and then reading them — makes it difficult to spot copy number variants. Gigantic mutations can thus hide in plain sight. Only now are scientists developing new methods to see them.
I asked several researchers to use those methods to scan my genome for copy number variants. At Yale, for example, Mark Gerstein and his team identified over 740,000 short insertions and deletions, known as indels. I also have about 2,000 big duplications and deletions in my genome, measuring over 2,000 bases. In one part of my genome, I have a duplication of 122,000 bases in a row. In another part, I’m missing a stretch of 163,000 bases.
“That’s a chunk,” Gerstein murmured when we looked over the data. “That’s a serious chunk.”
Cataloging copy number variants is just the first step toward understanding them. Sometimes they are a benefit. Sometimes they are a burden. And sometimes, incredibly enough, these radical changes to our genomes make no apparent difference at all.
When I asked Bob Handsaker, a biologist at the Broad Institute, to take a look at my genome, he paid particular attention to some of the genes he’s been studying recently. One of them makes an enzyme called amylase in saliva that helps break down starchy foods as we chew them.
The amylase gene has undergone a curious evolution in the past few million years. Our primate relatives only carry a single amylase gene, but we humans carry a number of copies of the gene — as many as 17 in some people.
Some scientists have argued that the evolution of extra amylase genes was an important step in the origin of humans. It’s possible our ancestors used their amylase enzymes to break down food more efficiently. The extra energy liberated by our extra amylase could then fuel our calorie-hungry brains.
With that hypothesis in mind, I was disappointed to learn from Handsaker that I came in well below average with only four amylase genes. I couldn’t help but wonder if I do a lousy job of drawing energy out of my morning oatmeal, starving my poor brain in the process.
Handsaker cheered me up with some science. It turns out that when it comes to amylase, humans range widely. Europeans typically have between six to eight genes, Handsaker and his colleagues have found, but some people have as few as two. “We don’t know why people have such different numbers,” said Handsaker.
It’s also not clear if having more amylase genes is good for your health. In 2014 and 2015, three teams of scientists reported that having fewer copies of amylase genes puts people at greater risk of obesity. Some researchers speculated that making less of the enzyme would make people not taste their food as much as others and eat more.
But when Handsaker and his colleagues looked more closely at amylase in larger groups of people, they failed to find a connection. For now, it looks as if my meager amylase gene count doesn’t have much impact on my life.
But Handsaker wasn’t going to let me off that easy. He had some disappointing news about another gene, which encodes a protein called haptoglobin. Haptoglobin helps control the chemistry of blood, grabbing molecules like hemoglobin and cholesterol and pulling them out of circulation.
In some people, the haptoglobin gene measures about 6,400 bases long. But other people carry a version from which 1,700 bases have been cut out. The short version of the gene makes a short version of the protein, which behaves differently than the long one. When long haptoglobin proteins bump into each other, they snap together in groups of four, forming an X shape. When short haptoglobins bump into each other, they snap together in pairs, forming a dumbbell-like cluster.
Handsaker and his colleagues discovered that when the long proteins form an X shape, they remove cholesterol slowly. The dumbbell shape formed by short proteins clears cholesterol faster. As a result, people who carry the long gene have higher cholesterol levels than people with the short gene.
This discovery, which Handsaker and his colleagues reported in February, marks one of the first times that scientists have been able to uncover the chain from a copy number variant to a protein’s structures to its impacts on health. “That’s a big thing to be able to get to the cause,” said Handsaker. “It gives a deeper understanding of what’s going on.”
As for me, Handsaker had some bad news. “You’ve got two copies of the slow kind,” Handsaker told me. “All other things being equal, you’re predicted to have more cholesterol in the blood. It’s just a roll of the dice, and you ended up a little worse on the cholesterol front.”
Handsaker and other researchers are embarking on new studies of copy number variants, looking for other effects. Sometimes these mutations alter how proteins work, as in the case of haptoglobin. Some deletions make entire genes vanish. Some duplications can give us extra copies of genes, which can raise the levels of proteins in our bodies.
“It’s doing something in your body,” said Handsaker. “Whether it’s a big effect or a small effect on your health, we don’t know. But as our tools improve, we’re beginning to find things.”
hen the conversation turns to genes, it usually gets scary: mutations, risks, diseases, death.
But it doesn’t have to be that way. Scientists know that some genetic variants actually protect us from getting sick.
Until recently, however, researchers haven’t looked for them very much. No one came to a geneticist demanding to know why they weren’t getting cancer or arthritis.
“I’m guessing there are hundreds of protective variants, maybe thousands,” said Eric Topol of Scripps Translational Science Institute, a research center that turns basic research into useful medicine. “We only know about a tiny fraction of them.”
I asked Topol how I could scan my genome for these protective variants. He pointed me to his colleague at Scripps, Ali Torkamani. But Topol warned me I might be disappointed. Torkamani might not find anything at all.
Torkamani was game to take a look, and after a few days of analysis, he was ready to talk. I gave him a call, and he explained he had come up with a list of 70 variants in my genome that looked like they might be protective. But after he took a close look at the evidence, he found most of it squishy. In the end, he felt solidly confident about only one variant.
When I found out about what that one variant does, I was very grateful to carry it: It protects me against an autoimmune disorder called Crohn’s disease.
Some 700,000 people in the United States suffer from this lifelong disease, which causes the digestive tract to become inflamed. People with Crohn’s disease may suffer persistent diarrhea and abdominal pain, and they may risk dangerous weight loss. There’s no cure for Crohn’s disease; the best that doctors can offer is medications that cool the immune system down and a diet that doesn’t trigger a new flare-up. Many people with the disorder eventually need surgery to have diseased pieces of the gut removed.
About a decade ago, Richard Duerr, a geneticist at the University of Pittsburgh, and his colleagues went on a search for genetic variants that raised the risk of Crohn’s. They found several variants that were unusually common in people with the disease. They were all in a single immune system gene called IL23R.
But the scientists also made an unexpected discovery: Another variant in IL23R was unusually uncommon in people with Crohn’s disease. Subsequent studies confirmed that it strongly reduces the risk of getting Crohn’s disease.
About 8 percent of people of European descent have that particular variant, known as rs11209026.
Torkamani informed me I am part of that minority.
Since Duerr’s study, other scientists have discovered even more good things about rs11209026. It also protects people against other diseases that involve the immune system, such as psoriasis and ankylosing spondylitis, in which inflammation causes vertebrae to fuse together in the spine.
As the evidence for the protective power of my variant grew, a number of researchers set out to discover where that power came from. They took a careful look at the gene it alters, IL23R. That gene encodes a protein that decorates immune cells.
Normally, the IL23R protein functions as a kind of radio antenna. It picks up certain molecular signals released by other immune cells and relays them into the cell’s interior. The cell then responds by turning on genes that trigger inflammation. That inflammation can be a valuable weapon against infections — if it’s produced in the right situations. If it’s triggered by false alarms, however, it can cause chronic damage, like Crohn’s disease.
Grant Gallagher, a scientist at the Institute for Biomarker Research in Hamilton, N.J., and his colleagues, have found that the rs11209026 variant has a drastic effect on IL23R. The entire protein slips out of the cell and floats away. It can still pick up signals, but it can’t relay them onward. As a result, Gallagher proposes, my variant breaks the chain of communication and stops the immune system from producing inflammation.
After Torkamani’s news, I found myself developing odd feelings about my variant. It’s hard to put those feelings into words. Maybe I felt smug, discovering I was part of a small fortunate club. Maybe I felt grateful, thinking of people I know who have to cope with decades of struggles brought on by Crohn’s disease.
Torkamani tempered my fondness for my variant with a warning. It’s entirely possible that future studies will reveal that ripping out my IL23R antennae puts me at risk of other diseases. “It may be a double-edged sword,” he warned. “You have to wonder, what are you sacrificing for this protection?”
It also turns out that I have other genetic variants that may actually increase my risk of Crohn’s disease.
Scientists still know so little about how harmful and protective variants combine to affect our health that they won’t become a part of standard medical care any time soon. “It’s hard to see how you’d use them for clinical advice,” said Torkamani. “I can’t see a physician saying, ‘You’re protected from coronary heart disease — eat all the hamburgers you want.’”
The greatest value of a variant like rs11209026 is not the protection it gives me and others like me, however. By studying it, scientists can learn about the complete biological picture of a disorder like Crohn’s disease. “It opens up the biology,” said Topol.
And once scientists understand a disease, they can test out treatments to cure it.
A number of pharmaceutical companies have already been trying to turn rs11209026 into a drug for years. They’ve designed drugs for Crohn’s disease and psoriasis that mimic the effects of the variant, breaking the inflammation communication relay in a similar way. Last year, the FDA approved the anti-psoriasis drug Cosentyx, made by Novartis, and other companies are getting promising results with other drugs now in human trials.
That being said, protective variants don’t always deliver slam dunks. Last year, Amgen abruptly canceled its own ambitious psoriasis drug program because a number of volunteers in the trial reported having suicidal thoughts. The double-edged sword of protection may cut too deep.
Scientists are now ramping up their hunt for protective variants. Torkamani, Topol, and their colleagues are studying more than 1,300 people who are over 80 years old but have no chronic diseases and don’t take chronic medications. They launched the study to look for genetic variants that helped their subjects enjoy a healthy old age.
In April, they reported in Cell a handful of genetic variants that were suspiciously common. Intriguingly, many of the genes that contain these variants make proteins in the brain — where they might shield old people from cognitive decline. For now, their results are only suggestive; scientists will have to look at many more healthy old people to replicate the study.
Torkamani checked my genome for the variants they’ve linked to healthy aging. He didn’t find any. I’m trying not to take that news too hard, though. I know that I may very well have my own collection of rare genetic variants that promote a healthy old age. And if I don’t, there’s always a chance that a drug made from one of Topol and Torkamani’s discoveries will help me out in my later years.
(Some of the scientists who analyzed my genome kindly provided their technical results, which you can see here.)
Correction: An earlier version of this story misspelled Ali Torkamani’s first name.