Researchers Comb Through Millions of Genetic Variants to Find Disease Risk

Your disease risk is hiding in your DNA, and doctors may soon be able to find it.

A study published today in Nature Genetics shows that it is possible to predict whether a person is at a high risk of developing a disease based on millions of changes across their genome. This genetic risk, quantified by a “polygenic score,” can be used by doctors to decide whether the patient needs preventative measures.

“It’s good to find people at high risk, but it’s even better to be able to do something about that risk,” said Sekar Kathiresan, director of the Center for Genomic Medicine at Massachusetts General Hospital, director of the Cardiovascular Disease Initiative at the Broad Institute, and professor of medicine at Harvard Medical School.

Genetic risk, in this case, is quantified by a “polygenic score,” and can be used by doctors to decide whether the patient needs preventative measures.

Kathiresan, the study’s senior author, speaks from experience. He is a practicing cardiologist and sees patients with a high familial risk for heart attacks. That’s where he first got the idea to look for heart attack risk in the genes.

Traditionally, doctors have thought of high blood pressure and high cholesterol as two of the main risk factors for heart attacks. The role of genetics has been limited to a mutation in a receptor involved in regulating cholesterol levels in the blood. But a prior study found that, among people who suffered heart attacks in their 30s and 40s, only around 2% had a mutation in the gene encoding the receptor that had been at the center of heart attack research for decades. This made Kathiresan wonder: What about the other 98%?

The answer, he realized, might lie not just in one gene, but dispersed across the entire genome: millions of changes to the genome that each contribute a tiny piece of the overall risk for the disease.

His team of researchers used the results from previous genome-wide association studies, which identify specific genetic changes linked to diseases. They designed polygenic scores that add up whether a person has each of the genetic changes, weighted by how much that specific change incrementally increases a person’s disease risk. They designed scores for five common diseases: coronary artery disease, atrial fibrillation, type 2 diabetes, inflammatory bowel disease, and breast cancer.

To figure out whether the scores were actually meaningful, the researchers took them for a test drive. They calculated people’s polygenic score for each disease and compared it to whether they actually had the disease in question. Ideally, people with a high polygenic score should be more likely to have the disease. And that’s exactly what they found.

For coronary artery disease, for example, the people with the top 8% of polygenic scores were over three times more likely to have the disease. They saw similar trends in the other diseases. People with the top 6.1%, 3.5%, 3.2%, and 1.5% of scores were over three times more likely to have atrial fibrillation, type 2 diabetes, inflammatory bowel disease, and breast cancer, respectively.

This isn’t the first time that researchers have tried to assemble an estimate of disease risk from scattered genetic changes, but previous attempts never quite identified enough risk to be useful in the clinic—you wouldn’t be too worried to find that your risk of developing a disease was 1.1 times higher than your neighbor, right? But increasing your risk 3-fold is more alarming.

“In my opinion, this is very exciting work,” said Ali Torkamani, director of Genome Informatics at the Scripps Research Translational Institute and associate professor of integrative structural and computational Biology at Scripps Research, who was not involved in the study. “The importance here is that there is little debate about the clinical utility of monogenic mutations, but a great deal of controversy around the clinical utility of polygenic risk. The findings add another strong piece of evidence that polygenic risk should be considered for use in clinical practice.”

Polygenic scores are even effective in people who may not have the traditional symptoms, but are still at high risk of developing a disease. And most importantly, polygenic risk is modifiable, Kathiresan said, and making lifestyle changes can cut the risk the half. If a patient has a high polygenic score for coronary artery disease, her doctor can prescribe statins preemptively. Or, if she has a high polygenic score for breast cancer, she can get a preventative mammogram.

The method isn’t perfect yet. A high polygenic score isn’t a guarantee that you will develop a disease, or a mandate to seek out preventative treatment.

“We have to think through the clinical implication for each disease and think about value of intervening knowing the risk,” Kathiresan said.

Its largest blind spot is a crucial one: people who aren’t white. This bias can be traced back to the initial source of the data. Since the mutations that go into the calculation come from genome-wide association studies done on white subjects, it’s not surprising that they work as better predictors in white patients. The same team of researchers is working on follow-up studies to design polygenic scores that are better predictors in people of east Asian, south Asian, African, and Hispanic ancestry.

Kathiresan’s vision for polygenic scores is one where they’re just one more common piece of medical information. Since the risk is determined at birth, he sees people getting genetic testing in early adulthood to generate a “report card” of their risk of various common diseases.

“In the future, people will know their polygenic risk for heart attack like they know their cholesterol now,” he said.

Calculating the polygenic scores requires knowing whether an individual has specific mutations. But that’s easier than it sounds. You can get this information from direct-to-consumer genetic testing services like 23andMe or Ancestry. The researchers are currently working on making their score calculator available to people who already have these genetic testing results on hand.

But they’re also in talks with companies to develop lab tests that can make these genetic tests available to everyone as part of the standard arsenal of diagnostic information available to health care providers. So, what might a doctor’s appointment of the future look like? The usual patient interview, lab tests, imaging—and now, polygenic scores.

“Polygenic risk scores are not currently used in the traditional clinical workflow,” Torkamani said. “They should be considered anytime a preventative medical decision is being made.”