How to test everyone for the coronavirus
“It’s my first global pandemic. How about you?” Jonathan Rothberg wanted to know.
Rothberg is a high-energy biotech entrepreneur who has been trapped in quarantine on his super-yacht, the Gene Machine, since mid-March, when we first reached him by phone. The creator of a fast DNA sequencing machine and, more recently, a revolutionary cheap ultrasound wand, Rothberg had been cruising in the Caribbean when the pandemic hit.
“I was on vacation and suddenly the world blew up,” he says. It didn’t take long for him to come up with a plan. He was walking by the small lab on his boat when he decided what to do. “I said, why don’t we figure out how to do this whole testing thing at home?” Rothberg recalls. Since then, he has been directing a team of about 60 at one of his companies, HomoDeus, to try to create an at-home genetic test for the virus.
Right now, gene tests—the most accurate kind—are run only in labs or on special hospital instruments. But Rothberg’s is one of several attempts to create a test that’s simple, foolproof, and cheap enough for anyone to do at home.
To Rothberg, it is obvious that tests need to become as ubiquitous as mobile phones if we want to get the world out of lockdown. “We can and will give everyone the ability to diagnose themselves now and in the future,” he said on Twitter, where he posts pictures from his boat and gives progress updates. “Any place. Any time.”
Who will do it?
The US badly fumbled the rollout of testing for the coronavirus. Tests remain limited, and so now the possibility that everyone can be tested, all the time, appears to ride mostly on private initiative, including ideas from a crew of clever entrepreneurs and top academic genome centers.
About half the globe’s population is under stay-at-home orders to stall the spread of the coronavirus. The next step, ending social isolation and restarting economies, depends on monitoring and tracking the virus. And for that, we may need massive testing, the kind done in parking lots and drive-throughs or at home, so people know if they are infected right away.
Proposals for getting back to a new normal, such as one advanced in Harvard Business Review, say doing so depends on being able to mitigate the infection. Key factors include avoiding overflow at hospitals, having masks for all workers and “the availability of sufficient testing capability” to identify anyone who is or has been, infected.
Exactly how much testing is that? Commercial labs and health departments are now testing around 140,000 people a day in the US, according to the COVID Tracking Project, and they can still enlarge their operations. But under a get-back-to-work plan we will need to go much, much bigger. While there isn’t a consensus on how big, some forecasts say we should be testing millions of people every day.
The White House created a new coronavirus task force, the Council to Reopen America, to restart the economy but has not detailed a plan around back-to-work testing. On April 11, Admiral Brett Giroir, the administration’s testing czar, said testing levels would rise in May to levels needed to support reopening, but he said testing everyone is “physically impossible” and a poor strategy.
That appears to leave the question of widespread, frequent, and repeated testing to the private sector. “Every employer and insurer wants it,” says Amitabh Chandra, a Harvard Business School professor and an author of the HBR back-to-work plan, but “everyone is thinking that someone else will do it.”
Mass testing, even if it’s expensive, could actually be one of the cheaper ways out of the crisis. MIT professor Erik Brynjolfsson says if tests cost $3 each, everyone in the US could be tested 10 times over six months for $9 billion. “That’s 0.3% of the cost of the recent bailout,” he tweeted.
Although there isn’t a clear federal or state plan for testing everyone, several research centers are on their way to establishing technologies they say can make it possible. “I think there are two ways to do it,” says Feng Zhang, a researcher at MIT’s Broad Institute. One is to “collect samples from many people and process them in one place” using ultra-fast genome sequencing machines. “Another way is to have a home test, a nucleic acid test that people can run in their homes, which is inexpensive and easy to do.”
Think of the first scheme as massive centralized testing. To drastically cut the cost of tests, Zhang’s lab and others are demonstrating how to look for the coronavirus using powerful next-generation sequencing machines, typically used to decode the billions of DNA letters in a human or animal genome.
The high-horsepower machines read out hundreds of millions of bits of DNA at once. Last week Zhang’s lab described an approach in which they could add individual molecular bar codes to patient samples, pool them, and then sequence them all at once. In theory, one sequencing instrument could test 100,000 samples at once, in about a day, for $7 each.
Jonathan Schmid-Burgk, who developed the method in Zhang’s lab, says the technology “could let you screen the whole population of the US every three months.”
One drawback of centralized schemes is there’s still going to be a delay in collecting and pooling samples and getting results back. Still, because of its large scale, it could be useful to create a covid-19 heat map that could alert governors when they need to reinstitute social distancing measures. “There is a question of how to scale it up, but the next-generation sequencing is already there. What you need is an academic center and a logistics partner,” says Schmid-Burgk, who is now in Germany and talking to partners there.
Sri Kosuri, founder and CEO of the biotech startup Octant, who is on leave from the University of California, Los Angeles, has posted plans for a mass testing system that he’s calling “SwabSeq,” with which one person could carry out 10,000 tests with about $10,000 worth of supplies. But like Zhang, he worries that his idea won’t be put to use, since the molecular biology teams aren’t in a position to figure out how to obtain 100,000 or a million swabs, or get them shipped to one location, much less tackle data entry or figure out how to return results to so many people.
“If you talk to the logistics people, they think theirs is the easy part but have little faith there is a technology that can scale to their capacity,” says Kosuri. “While the technologists say that the technology is the easy part and the logistics is hard. It’s a little bit chicken-and-egg. We should be betting on both and hoping to deliver, because the alternative is we both don’t do anything.”
I asked Kosuri if a tech hub like Boston or San Francisco couldn’t move forward on its own with a mass testing scheme, perhaps led by a university and backed by local officials. He said that might be what happens. “All that needs to happen is for one place to execute at scale, and then everyone will do it,” he says.
Testing at home
The other way to get massive testing under way is to make it decentralized—that is, let everyone test at home, whenever they need to. An at-home genetic test would let everyone make their own decisions. “If it’s cheap enough,” says Zhang, “people could do two tests a day, one in the morning and one in the evening. Then they can decide if they’re going to work the next day.”
“You want to get information back to people, so they can use that information to not spread it to other people,” he says.
One type of test that’s possible to do at home—with a finger prick—is called an antibody test. It looks for signs in your blood that you were already infected. These signs don’t develop for a week. By contrast, genetic tests look directly for the genetic material of the coronavirus. They are more accurate and can identify infected people even before they show any symptoms.
Right now, though, there is no genetic test of any kind for home use. That’s because most gene testing methods require several laboratory steps and a PCR machine to rapidly heat and cool a sample (a step to help copy or amplify the genetic material). But researchers including Zhang say they are close to demonstrating at-home technology for covid-19 using new types of chemistry that skip the heating and cooling cycles.
Zhang’s lab is best known for its role in developing the gene-editing tool CRISPR, which he and others have crafted into a new type of genetic diagnostic method (if a gene from the coronavirus is present, the CRISPR molecule cuts it and sets off a molecular alert). Zhang says he started working on a CRISPR-based coronavirus test in January “when I first heard about the virus,” and by Valentine’s Day had uploaded a CRISPR protocol for detecting the virus in a laboratory.
It’s a fairly simple test that takes only hour, but it still involves a mixing step and heating up the sample once, to about 65 °C, to break open the virus and release its DNA. Now the lab is trying to simplify the process even further. “The goal is one-step reaction: you run it, open the tube, put in a paper strip,” he says. “That is something that we have been focused on in the lab for a couple of months.”
The spit problem
Rothberg is in quarantine aboard the yacht with his children, including a daughter who takes immune-suppressing drugs for a genetic disease, putting her at high risk if she gets infected. “She absolutely can’t be exposed,” says Rothberg, “My wife said, ‘Nobody comes on or off the boat.’”
To create a home DNA test, Rothberg redirected the efforts of HomoDeus, which had been trying to develop gene therapies using special enzymes called recombinases, a type of gene-editing tool. The chemicals also offered a path to do home gene tests, but before the pandemic, Rothberg says, he didn’t pursue it because it’s hard to make money on diagnostic tests.
“I would go to Davos and every year think about doing it on a paper strip, but never got around to it,” he says. “We’ve all made huge mistakes by not acting, so I feel that I have to do what is morally right and make a home test.”
At a complex of research companies Rothberg started in Guilford, Connecticut, there are usually 1,000 people at work, but now only about 10 lab workers remain on site, with the rest remote or working in basement laboratories.
Since last week, the new chemistry developed by HomoDeus has entered tests at Yale University on swabs from covid-19 patients. The chemical components are mostly solved but performing the test still involves some moving and mixing of fluids. “We still have a process a 17-year-old could do, but not a 70-year-old,” he says.
Rothberg says there are already prototypes of a home device that would automate those steps—it looks like a thick credit card with chambers to store liquids in. Another version employs caps filled with chemicals that can be screwed onto a tube. He’s been tapping his network, he says, to get people to quickly prototype circuit boards and 3D-printed models.
Rothberg says there is still “huge technical risk” in the project, but it’s not whether the chemistry works; the problems are more practical ones. It’s easiest for someone at home to spit in a tube, but it’s unclear if that will yield enough of the virus to test for. “If they just ate a chicken sandwich it could throw things off,” Rothberg says.
He says instead he’s leaning on the Gates Foundation, which has been working to prove that a soft swab passed inside the nose can serve as the basis for either mail-in tests or at-home ones.
After the slow rollout of testing in the US, the Food and Drug Administration began offering “emergency” authorizations so companies could introduce new test formats much faster. Rothberg says he anticipates little resistance to an at-home coronavirus gene test and says the agency has been helping him. “I am pretty confident our problem isn’t regulatory; it’s technology,” he says.
So how soon will it be ready? He’s challenged his team to have 10,000 prototype kits made next month, along with an accompanying phone app. “This is like conquering smallpox,” he says. “It’s going to be a project.”