Latest News

 

By Aaron Krol 

June 22, 2015 | Before they sold cut-rate tests for breast and ovarian cancer, Elad Gil and Othman Laraki worked at Twitter. It’s one of those great Silicon Valley success stories: a small team, a good idea, a rapid acquisition. In 2008, the two tech developers left their jobs at Google to found Mixer Labs, a startup focused on geolocation — tools that tell app developers where their users are. Their platform got the attention of the executive team at Twitter, who scooped up Mixer Labs a year later. Gil and Laraki got to be Vice Presidents at one of the world’s largest social media companies, and Twitter got the tags on tweets that tell you what neighborhoods they were posted from.

But that was 2009, when Twitter was riding a new wave in the way we communicate. Now it’s just treading water. Gil and Laraki, successful angel investors with a knack for sensing the next tech trend to crest, have moved on from social media to genomics, where they expect their skills will be gladly received.

“If you think ahead to the future of genomics,” said Gil in an interview with Diagnostics World, “we think it morphs from a wet lab problem into a data problem. How do you deal with this information over a lifetime?”

His answer is Color Genomics, the genetic testing company of which he and Laraki are now CEO and President, respectively. Color’s first product, a test covering 19 genes that affect women’s risks of developing breast and ovarian cancer, is a forceful demonstration of how small a barrier the wet lab processes have become. By using the latest DNA sequencers, a high level of parallel testing, and some custom lab automation, Color is able to profitably sell its tests for $249 apiece — roughly one tenth the going price.

That’s not just a competitive edge. It’s fundamental to the idea of Color, which sells its tests online and charges customers directly, rather than negotiating with health insurers. Insurance companies are only obligated to pay for genetic testing for those patients who have some preexisting risk, like women with family histories of breast cancer. Men, who could pass on these mutations to their children but are much less likely to be directly affected, have even less access to testing.

“Half of the women who carry mutations of these genes don’t have the family history that would allow them to get tested,” says Gil. “This information can help everybody, not just the select few who can pay for it or have insurance coverage.”

Several of Color’s executives could have benefited from a service like this themselves. Laraki, for one, has lost family members to hereditary forms of cancer. Early warning about this risk might have helped his relatives prepare, and opened up better treatment options.

Laraki, like hundreds of thousands of other Americans, carries a mutation in a gene called BRCA2 — a mutation that causes cancer, early and predictably. That’s the kind of bold statement you won’t often hear from geneticists, who usually deal in incremental risks spread across dozens of genetic variants. But mutations to BRCA2, and the related BRCA1, are different animals. Women with certain variants in these genes, if they don’t take preventive measures like undergoing mastectomy, stand an 80% risk of developing breast cancer over their lifetimes.

At Color, Gil and Laraki are interested in the data and access problems that come with genes like BRCA, which in the lingo of genetics are highly “actionable.” Some mutations in these genes are still new to science; how do you interpret those? Where do you store the information? How can you connect patients with the health services they need to understand and act on their results?

Where it once seemed that these questions would be answered in the confines of hospitals and doctors’ offices, the rise of companies like Color suggests they may end up being settled by the free market. No one is watching over Color’s shoulders when it chooses which genetic variants to report to its customers, or what to tell them next. Operating under few formal regulations, and speaking directly to patients, Color has to decide for itself what standards to follow.

For a company that believes its brand of widely available genetic testing can save lives, that’s a heavy moral burden to take on.

A Different Standard 

Color Genomics only opened for business this April, but its founders had been thinking about the future of DNA testing for a long time. As an investor, Gil is always on the lookout for markets on the cusp of taking off, an instinct that has won him stakes in influential companies like AirBnB, Square, and Pinterest. “Genomics hasn't hit the mainstream hype cycle yet,” he wrote recently on his blog aimed at young tech entrepreneurs. “[But] in late 2015 or early 2016 it will emerge as a hot area of investment.”

Others in Silicon Valley reached the same conclusion a decade earlier. In 2006, the pioneer in this field was 23andMe, which launched with backing from Google and a plan to market affordable and broad-ranging genetic tests through the web. The tests, which included comprehensive health profiles covering conditions like cancer, heart disease, and diabetes, were a hit, reaching more than a million customers to date. 23andMe was one of the first companies to show how a smooth web platform and a perky marketing campaign could persuade people to see genomics as a tool for personal betterment, not just a vaguely disconcerting laboratory science.

Ultimately, however, 23andMe would not be the template for a new generation of genome services. In 2013, the FDA declared that 23andMe was selling medical tests without submitting to premarket approval — a violation that led to its health reports being stripped from the market.

Which is probably why, despite its popular success, 23andMe is emphatically not the model for Color.

“What 23andMe does is fundamentally different from what we’re doing,” says Gil. He sees all kinds of flaws in the way 23andMe crafted its health tests: they looked at just one particular kind of mutation, changes to single DNA letters. They looked at only a select few of those. They drew overly broad conclusions. “For BRCA1 and 2, I think they covered three alleles in total,” he explains. “If you look at these genes, there are literally thousands.”

 

Gil wants to hold his company to a different standard. If you buy a Color test, you will have every DNA letter of all 19 genes sequenced. If Color has never seen a particular mutation before, they’ll try to find it in the scientific literature and public databases, and figure out whether it’s harmful or benign. A custom software platform lets the company’s scientists quickly profile every genetic variant that shows up in every test.

“We do literature curation on all the variants,” Gil says. “We do in silico predictions. We look at population frequencies.” And standards set up by professional organizations, like the American College of Medical Genetics and the National Comprehensive Cancer Network, provide guidance for how Color classifies genetic variants — on a five-point scale from “benign” to “pathogenic” — and what medical advice the company offers when customers test positive.

But in the end, the most important difference between Color Genomics and 23andMe might be something that happens well before Color even gets your DNA sample — right after you enter your credit card information on the website.

At that moment, two boxes appear in your browser. The first one, checked by default, reads: “I’d like a physician designated by Color to order my testing.” The second reads: “My own physician has ordered Color for me.”

Any test you buy for yourself, from a pregnancy test to a box of glucose test strips, needs to go through a premarket approval process with the FDA. Physicians, though, are another story. When the regulations around medical tests were written, no one wanted to interfere with doctors’ ability to treat their patients as they saw fit. As a result, physicians are free to order almost any tests they like, except in a few states like New York and Maryland with more stringent requirements.

Color operates in 45 states, and has relationships with physicians in all of them prepared to order tests on behalf of customers. It’s a situation that puts the company outside the scope of current FDA regulations, just like a hospital lab running tests for a doctor through insurance.

Gil says his company probably would have worked with physicians whether or not the FDA had let businesses like 23andMe sell health tests directly to the public. For patients who have high-risk mutations, he says, it’s Color’s responsibility to offer professional help deciding what to do next. “Ensuring that they’re working with their doctor — or working with a doctor — on a plan moving forward is really important,” he says.

Color declined to share how often it provides physicians to order tests on behalf of its customers. Gil says a significant number of buyers either order through their own physicians, or at least make sure their usual care providers get copies of the results. Nonetheless, by adopting this business model, Color Genomics is testing the limits of what it means to sell direct to patients.

“As Many People as Possible” 

“It’s interesting timing,” observes Saskia Sanderson, a professor of genetics at the Icahn Institute of Mount Sinai. “Just a short while after 23andMe were stopped from selling their health information, Color Genomics and a few other companies are appearing with this kind of genetic testing. It kind of starts the conversation again.”

Sanderson’s research focuses on how patients respond to genomic testing. In one recent study, the Icahn Institute interviewed 35 people who were offered free whole genome sequencing, to learn what these people wanted to know about their genomes, what worries they had about the testing, and how they would deal with actionable results. “This is coming,” Sanderson says. “Personal genomic information simply is going to become more and more available. We have to focus on finding ways to do this right, and to do it right for as many people as possible.”

Her research suggests that there’s a high level of public interest in personal genomics, and that the biggest driver of that interest is the belief that DNA information can help people make important decisions about their health. That’s surely good news for Color and its counterparts — companies like Invitae of San Francisco, and Kailos Genetics of Huntsville, Alabama, which have begun selling low-cost genomic tests online through physician intermediaries. All three companies started offering these services in just the past three months.

“It’s good that they get a healthcare provider involved on some level,” says Sanderson. “But it raises questions about what the purpose of going through the healthcare provider is. In effect, the individual purchasing the test is still doing that through the website.”

There are a number of roles physicians could play in genomic testing. The more genetically savvy can be a first screen on results, making sure that reports seem accurate and actionable before they reach patients. Over the long term, doctors can also set up plans to respond to genetic risk factors — whether that means screening periodically for signs of cancer, or in the most extreme cases, arranging for preemptive surgeries.

Most doctors, however, don’t get a great deal of training in genomics. For that, we have genetic counselors, professionals whose job is to help people understand the implications of testing — from the risk of false positives and negatives, to the ways family members might be affected by the results.

“It’s very important to provide [counseling] as a service,” says Gil. “We’ve spent a lot of time with various physicians, clinicians, and genetic counselors to ensure that the type of content we provide to people reflects the type of information they thought was important to convey.” Color, like most of its counterparts, also offers a free session with a genetic counselor to all customers.

“Providing a genetic counselor to talk to the patient about their test results once is good,” says Sanderson, but there is a conflict of interest to navigate. “Genetic counseling is also about helping people make informed decisions about whether this is right for them. They need information up front that will genuinely help them make an informed choice. And that has to happen before they go ahead with the testing.”

For a company like Color, she adds, “that’s clearly a little bit difficult.”

More Access, More Questions 

We have no national plan for genomic testing in the United States, and with our fragmented healthcare system, we’re unlikely to get one. For better and for worse, companies like Color will soon be taking the lead in bringing genomic information to the general public.

“It’s going to become cheaper to sequence the whole genome than it is to do targeted genetic tests,” says Sanderson. “It simply is going to be cheap and useful in the future, and it’s going to be extremely widespread. I don’t think that is really in question.”

The more access we have to our DNA, the more questions will arise about what information is medically relevant  — and who’s going to decide. Even the fairly narrow tests that can be sold at out-of-pocket prices today have to wrangle with these considerations. Kailos Genetics, for instance, sells tests related to drug response, for medications like blood thinners and antidepressants whose effects vary based on patients’ genetics. That’s a kind of testing that many hospitals have been hesitant to adopt, in part out of fear that our knowledge about drug-gene relationships may be too incomplete to be useful.

In breast and ovarian cancer, where BRCA mutations have an outsize presence, Gil and Laraki have chosen one of the least contentious fields in which to begin a genomic testing business. But even here, there is room for controversy.

 

“A mutation being found in a BRCA1 or 2 gene is clearly very different from a mutation being found in one of the other 17 genes [in a Color test],” says Sanderson. That’s not because the relationship between these other genes and cancer is in dispute; several, like CHEK2 and TP53, have been known to cancer researchers for decades. But their effects are not as strong or predictable as those of BRCA mutations, and that poses real problems for patients who may learn they have a heightened, but not overwhelming, future risk of cancer.

“We’ve purposefully tried to exclude genes if the impact they have to a person’s risk is very marginal or low,” says Gil, noting that Color’s 19-gene panel is roughly equivalent to similar tests paid for by insurance. Color also takes care not to treat every mutation the same way. The company draws heavily from guidelines written by the National Comprehensive Cancer Network, which recommends different kinds of screening programs and interventions for patients with different genetic profiles. (Although one piece of advice is left out: the NCCN does not recommend screening most patients for these mutations in the first place, absent some compelling family history.)

Gil is no naïf to genomics. Before taking his talents to Silicon Valley, he received a PhD in biology from MIT. As a doctoral candidate, he studied a gene called PTEN that was one of the first to be recognized as playing a role in cancer — and which is now included in his company’s tests. When Gil and Laraki formed Color Genomics, they also brought in Taylor Sittler, a breast cancer expert from the University of California, San Francisco, as a co-founder, and convinced Mary-Claire King, the discoverer of BRCA mutations, to serve as a scientific advisor.

These professionals realize that their business is a delicate one. Gil even acknowledges some concern that other companies might start selling genomic tests without following the same voluntary standards Color has; as long they can find physicians to order tests, there’s little stopping them. But offering more people access to the medical power of the genome, he feels, is an important calling.

“The focus of the company is really to democratize access to genetic testing,” he says. “And to do it in a responsible and in a thoughtful manner.”

Like other transformations in 21^st century medicine, from electronic health records to high-priced biological drugs, widespread DNA testing brings a dilemma with every promise. One episode from Sanderson’s research stands out especially vividly in her mind. Her team originally wanted to recruit poor and uninsured volunteers for their studies, knowing that most research on genomic testing had drawn disproportionately from the wealthy. But when they went to a community clinic to recruit patients, her team was rebuffed.

“The people running the clinic were very concerned about the costs of follow-up care: the referrals if something was found,” Sanderson remembers. Worried that patients unable to pay for care would have trouble coping with their cancer risks, the clinic decided it would be irresponsible to test them in the first place. “If we really care about giving access to people who don’t have a lot of money, then we have to take into account the fact that the follow-up care is going to cost money. I don’t think we have the answers as to the best ways to do this at scale.”

Still, Sanderson is an optimist. “None of this is to say we should not be doing this,” she concludes. “And I’m not sure I feel that more regulation is necessarily the answer. I’m not sure saying that companies shouldn’t be doing this is necessarily the answer.” Instead, her concerns simply underscore a gap between what it takes to make a profit selling genomic tests, and what it takes to make sure those tests are a positive force for the buyer.

Meanwhile, Color Genomics is taking its mandate to expand access very seriously. Customers are urged to donate to a program called Every Woman, which will let Color give its tests for free to people who can’t afford the $249 price tag.

It’s a fine idea, even a noble one. And like everything else in this peculiar new business, it raises questions all its own.