"We're driving a medical revolution," says Venter of his latest startup, Human Longevity Inc., or HLI. "We have sequencing that's better than anybody else in the world. We have the most accurate data."
Venter, one of the best known scientists of the 21st century for his role in mapping the human genome, has been striking that kind of lofty note a lot lately, and investors are loving it. The four-year-old company has raised as much as $500 million, including $200 million this year, giving it a valuation of about $1.9 billion, according to data provider PitchBook. Plans are in the works for yet another round of venture capital before ultimately going public, according to Venter.
The problem is Venter's promises are ringing hollow for a growing chorus of critics. Conversations with more than a dozen current and former employees, customers and medical professionals depict a company that may prove unable to keep up with its founder's ambitions. Some doctors contend that such comprehensive testing isn't particularly useful, and even those who do think so say rivals may outpace Venter. Competing government-backed efforts in the U.S. and the U.K., for instance, threaten to overtake HLI in the race to collect massive amounts of the population's genetic and clinical data, a key proposition for the company's business success.
"Is HLI going to stay far enough ahead that these projects don't make it obsolete?" says Daniel MacArthur, co-director of medical and population genetics at the Broad Institute of Harvard University and the Massachusetts Institute of Technology.
Venter has no doubt about that. He says 40 percent of the 1,000 or so customers who have taken the test, called the Health Nucleus, have learned they have a "serious disease." Some come from China, through a Shanghai-based medical tourism startup called Quantum Clinics, which partnered with HLI in 2016. Wealthy Chinese, willing to pay out of pocket, are attracted to HLI's white-glove service and attentive care, according to Chief Executive Officer Lu Yi, who said Quantum will send about 50 Chinese patients for testing over the following year.
Yet as Venter touts that 40 percent disease-discovery rate, a paper posted online earlier this year by HLI scientists said age-related diseases requiring "prompt" medical attention have been found in just 8 percent of participants. Venter says the difference in numbers is because "it depends on what you count." He includes findings like pre-diabetes that would lead a person to change his or her lifestyle.
Some doctors doubt the need for such extensive tests. "I don't like the idea of selling a product by scaring people," says Ethan Weiss, a cardiologist at the University of California San Francisco. Weiss looked over a list of conditions that Venter said had been caught by the test. He took issue with cases of dilated left atrium, which can be a predictor of cardiac risks. "Maybe your risk of developing atrial fibrillation over a lifetime is higher, but it's not clear today that this is valuable information. Your heart's not going to explode."
For a customer who discovers a dangerous, unknown tumor, there's no question that the test would be priceless. But for those who don't receive clear-cut findings, the volume of data generated by the test can be overwhelming.
"I got a 400-page report and no one knew what to do with it," said Emily Melton, a partner at the venture firm DFJ, who took the test and shared the results with her doctors. DFJ has invested in HLI. Steve Jurvetson, who led DFJ's participation, declined to comment.
Venter says the problem isn't with his company's reports, but rather with the medical community, which is untrained and can't keep up with his technology.
Skeptics doubt that Venter can get there before massive, government-backed efforts, which are embarking on similar projects. "Over the next three years, the U.K. Biobank will sequence and make public exome or genome data from 500,000 well-phenotyped people. And over the same time frame, the U.S. 'All of Us' program will hopefully sequence more than 1 million," MacArthur says.
To ramp up its customer base, HLI has resorted to slashing prices. In May, it created a $7,500 option with reduced services, then added a $4,900 offering for just the genome and whole body MRI. The website has been offering a limited time promotion of $2,500. Venter says he's working on signing contracts with large employers and expects to have "literally tens of thousands" of customers through those corporate partnerships.
As the race heats up, Venter's bolder statements have irked some of the company's current and former employees, who agreed to speak on the condition of anonymity. Two former employees described chaotic early days where the company was working on as many as 10 different projects, from forensics to a search engine for genomic information. Multiple cancer projects were underway at one point, including a service that would sequence a patients' tumor tissue and compare it to normal tissue. There was also a short-lived excursion into pediatric rare diseases.
That's "a fair criticism," said Venter, who stepped aside as CEO in January when HLI hired Cynthia Collins. He remains executive chairman of the board and speaks for the company. He added that some trial and error was necessary to figure out where the startup would be most successful. "We're concentrating on our uniqueness and probably have eliminated 100 or so of our diverse test projects."
The potential gap between Venter's unbridled enthusiasm and his startup's actual capabilities spilled into the public this month when a paper was published in the journal Proceedings of the National Academy of Sciences. In it, HLI scientists said they could use a DNA sample to accurately identify an individual in a line-up of 10 people about 80 percent of the time. Most of the identification power came from predicting the gender, race and age. The scientists also used the information to render a facial image.
Jason Piper, one of the 28 authors, who had left HLI prior to the study's publication, took to Twitter to criticize the paper. "Don't get me started about how terrible those faces are," he said, arguing that the images generated by HLI's algorithms were simply the average face of a given ethnicity, so naturally, many people would look like the average.
Two other authors of the paper, who asked not to be named because they weren't authorized to speak for the company, said the shortcomings of facial predictions were due to the small sample size of 1,000 participants and the images would improve with more data.
Meanwhile, Venter was telling media outlets, "When you have the whole genome, we can predict a photograph." That led to headlines like this one, on San Diego's KPBS website, trumpeting: "These San Diego Scientists Can Predict How You Look Using Only Your Anonymous DNA."
When told about Venter's statement, another author of the study paused. "Well, there's marketing and there's the truth," he said. "You need to separate those things."