Skip to Main Content

PEARL RIVER, N.Y. — Delta was tearing its way through India and, as the death toll climbed higher, so did the anxiety inside Pfizer’s lab north of New York City. The Covid-19 variant was threatening to undo their work from the past breakneck year, by rendering their acclaimed vaccine ineffective.

Hundreds of millions of doses have been injected worldwide, but at the Pearl River research center, the pace has not let up. A team of “variant hunters,” as they call themselves, race to track changes in the fast-mutating SARS-CoV-2. A “virus farmer” grows the latest variants so researchers can test how they fare against the vaccine. And a colleague known as the “graphing unicorn” converts the data into intelligible results overnight.

The scientist leading all this work, Phil Dormitzer, was among the first to open the email bearing results of tests on how well Pfizer’s shot worked against Delta. For a heart-stopping moment, he thought the vaccine was indeed less protective against this wildfire of a variant. Then he looked again.


“I realized that no, the one that was spreading was not the one with reduced neutralization,” he said. His team had tested two virus strains that emerged in India around the same time, and only one showed a drop in potency. Delta, which was fast becoming dominant, was efficiently wiped out by the vaccine.

An exhale reverberated around Pfizer’s team that June day. Kena Swanson, senior director of viral vaccines, mimed wiping her brow with relief. “So far everything has looked actually quite good,” she said.

Delta still has ways to cause more breakthrough infections in vaccinated people: It spreads more efficiently, has a shorter incubation period, and produces a higher viral load in those infected. But the test results showed the vaccine itself was working well.


Phil Dormitzer, chief scientific officer for viral vaccines, outside the Pfizer Pearl River labs. Desiree Rios for STAT

STAT was granted a rare look inside Pfizer’s Pearl River research center, which has remained a place of frenetic activity for its 900 workers through the pandemic. The long, squat red brick buildings have operated as a laboratory for more than a century and played a role in past international emergencies, producing penicillin and typhus vaccines in World War II and the major oral polio vaccine in the ’60s. The site is now adapting to the current crisis, as unprecedented numbers of patient specimens crowd its loading bays. Inside the white, fluorescent-lit corridors, Pfizer built a high-security laboratory with enough safety protocols and air locks to create copies of the variants and contain SARS-2.

Monitoring the virus so closely has yielded some surprises, including calling into question old assumptions about vaccines: Recipients have protection before the vaccines have generated a significant antibody response, for example, a finding that is leading researchers to devise new methods of evaluating future vaccine candidates.

The work is exhausting, but the researchers cannot afford to pause. “My goal is to go to bed before the sun rises,” said Kristin Tompkins, associate director of vaccine research and development, and the first person at Pfizer to receive data on how each new variant responds to the vaccine.

On the day STAT visited this month, a few waves of her red hair were out of place, her immaculate look fraying slightly. “A lot of times that doesn’t happen.”

Pfizer’s discoveries are inextricably linked with the global political response to the pandemic. The day before STAT’s visit to the lab, the Biden administration declared all Americans should get boosters, a call that followed the company’s announcement that a third dose would likely be needed six months to a year after the first.

The controversy around this decision hasn’t escaped Pfizer’s own scientists. “There was talk this morning on TV about boosters. But I would prefer to see more people get vaccinated,” said Vidia Roopchand, principal scientist for viral vaccines. This makes sense from an epidemiology standpoint, he added: “It means fewer opportunities for the virus to mutate.”

Pfizer’s vaccine declines in protectiveness over time but remains more than 90% effective against severe disease six months on. Roopchand’s colleagues were equivocal on whether it makes sense to use boosters now. The extra doses will either be early or late, said Swanson and Dormitzer, and it’s better to be the former. “In terms of the exact timing, that’s really a decision for public health authorities, not for Pfizer to make actually,” said Dormitzer, Pfizer’s chief scientific officer for viral vaccines.

Booster shots don’t prevent the much-needed wider vaccinations, he added: “I don’t think every dose given in the U.S. is taking a dose away from somebody else around the world. We’re producing vaccines around the world on a continuous basis in large numbers.”

Just as Pfizer’s work informs policy, so do government choices shape their research. Swanson leads the team of “variant hunters,” and her work will not ease until the vast majority of the world has been vaccinated. “That’s the real goal. If you can do that, the opportunity for these kinds of things are much reduced,” she said.

She has an air of calm despite the intensity of her work. There are endless emerging variants, but Swanson picks the most worrying handful to pay particular attention to: those that are spreading internationally, are increasingly dominant, and are linked with higher death rates. The countries buying the vaccines also influence the choice of variants to study. “This also is a product, and so it’s also responding to customers,” said Dormitzer. “And the main customers are governments — what do they want to see as well?”

Researchers must pass any equipment going to the biosafety level 3 lab at Pearl River through this air lock. Desiree Rios for STAT

The real-world impact of variants takes months to play out, but Pfizer gets an early look at the threat by running lab tests on each variant.

Isolating and transporting each particular variant from around the world is too much of a hassle. Instead, Pfizer scientists grow the variants themselves, using information from a database called GISAID, which contains virus sequencing data uploaded by scientists and physicians worldwide. The genetic sequence acts as an instruction manual. Swanson’s team downloads the sequences attributed to a particular variant, and runs an algorithm to identify the key mutations. Next, they synthesize the new spike protein, which is what SARS-CoV-2 uses to latch onto and enter human cells.

Most other research centers insert the spike protein into a different, non-pathogenic virus, creating what’s called a pseudovirus, which requires a lower lab security level, but is also less accurate. And so Pfizer’s scientists use another method: The company has samples of the virus from the first Covid-19 case in the U.S., back in January 2020. To create a new variant, its scientists insert the new spike protein into the old SARS-2 virus.

“Our system is very different from a lot of other labs, and I think we’re the only one in the world doing it this way,” said Pei-Yong Shi, a molecular biology professor at the University of Texas Medical Branch who’s working closely with Pfizer to study the variants.

Pre-pandemic, Pfizer’s Pearl River research center didn’t have a biosafety level 3 lab required to grow Covid-19 — with an air lock and scientists clothed in hooded white suits, gloves, impervious gowns, and using an air-purifying respirator. (Such a lab didn’t open until March 2021.) And so the spike protein is packed in ice and flown to Texas, where Shi inserts it into the existing SARS-2 virus to create the new variant.

This virus is mixed with serum from the blood of vaccinated people, which contains antibodies their immune systems made in response to the vaccine. Different dilutions of the serum are used in a neutralization assay to show what quantity of serum it takes to inactivate the virus. The more serum that’s required, the less effective the vaccine is against that particular strain.

The test only points to the antibody response, not other aspects of the immune system that also protect against viruses, such as the T cell response. “We certainly take it with a grain of salt,” said Dormitzer. But while they’re far from perfect, the results are still a useful indicator.

Vidia Roopchand, principal scientist for viral vaccines, looks at cells under a microscope. Desiree Rios for STAT

There’s no perfect recipe to grow a virus. It takes warmth and glucose, said Roopchand, who oversees the process at Pearl River. It also takes a certain vibe.

“We treat viruses with a lot of respect. Not fear — respect!” he said, with a delighted laugh. “If you fear them, then they probably won’t grow.” Viruses are a part of nature, he added. They have every right to be here.

Roopchand has plenty of good vibes. He bounces around his laboratory, gesticulating rapidly, with the energy of a child on a sugar high rather than a scientist who’s worked nonstop throughout the pandemic.

“We treat viruses with a lot of respect. Not fear — respect! If you fear them, then they probably won’t grow.”

Vidia Roopchand, Pfizer principal scientist for viral vaccines

Last Christmas, he was due to take his first break since March 2020. The day before his vacation, Dormitzer said they needed to assess new threats. Variants that originated in the United Kingdom, Brazil, and South Africa were all spreading quickly, derailing gatherings and intensifying fears.

Roopchand, who refers to himself as a “vaccine farmer,” canceled his plans and began inserting spike protein from the variants into non-pathogenic viruses. It felt like his responsibility, said Roopchand, who was born in Guyana. “I’m from the developing world and I’ve seen what infectious disease can do,” he said.

And so, on Christmas Day, he was in the lab making sure the cells were dividing properly. “You have to have happy cells,” he said. “You talk to it.”

The conversation paid off. Roopchand developed all three strains, then handed them to his colleagues for testing. Swanson’s team would get the results back on Roopchand’s one day off, on the Sunday between Christmas and New Year’s. He couldn’t relax.

Roopchand tried to exercise as a way to wind down, but it didn’t help. As soon as he came home, his wife wanted to know why he didn’t have his post-workout glow. “She knows me well, we’ve been married 20 years. One thing about science is you carry it with you, and your body language betrays you.”

Finally, late in the afternoon, he got the results by email. It was a good Christmas, in the end, because he knew their vaccine worked.

Packages of serum samples from participants in studies of the Covid-19 vaccine arrive at the Pearl River facility. Desiree Rios for STAT
Robert Lerch (right), an associate scientist, waits as employees of Clancy Relocation & Logistics unload a package of serum samples onto dry ice. Desiree Rios for STAT

Serum from the blood of vaccinated participants is easier to get than the latest SARS-2 variants. Every day, a hundred boxes, containing 3,500 blood and nasal samples, arrive at Pearl River from sites conducting clinical studies testing the long-term efficacy of the vaccine and booster shots, as well as from vaccine trials in children and pregnant people.

Pre-pandemic, scientists unpacked boxes in a lab containing a giant robotic freezer the size of 20 standard freezers. Once Covid-19 trials were running, they ran out of both space and the ability to manage so many deliveries.

“Having our scientists lift and dump out heavy boxes of dry ice a hundred times a day wasn’t the most efficient way to go about things,” said sample management specialist Jacob Stass. “I get a bad back from just getting out of my chair too quickly.”

Last summer, they switched strategies to have a team that normally moves heavy equipment unpack the boxes in the receiving dock. By late morning on the Tuesday a STAT reporter visited, dozens of large cardboard boxes sat by the door, each containing dry ice and a smaller box with specimens inside. As soon as they were opened, each specimen was moved onto a table containing dry ice, where it was verified and signed before being moved to the main freezer.

Some of these specimens are sent on to Texas, where Shi tests them against his newly grown variants. Faced with travel disruptions during the pandemic, Pfizer started using its corporate jets to transport the samples as quickly as possible.

“The logistics of getting materials back and forth was proving very difficult,” said Steve Bjornson, vice president and chief operating officer of vaccine research and development. The fleet was typically used to fly executives to priority meetings but, with more business taking place on Zoom, they were increasingly available and Pfizer chief executive Albert Bourla suggested the planes could be used to fly biologics instead.

Once the spike is generated, the virus is grown, and the neutralization assays have developed, the final stage is to analyze the results. Tompkins, the “graphing unicorn,” has this job. “I get late-night raw data and then, magically in the morning, they have beautiful graphs to interpret,” she said. Shi often sends her data late in the evening at 10 p.m. or sometimes even 1 a.m.

The updates are exhausting but exhilarating, said Tompkins. “Every time I get that clinical data, it’s a thrill just to see how awesome our vaccine is,” she said. “It’s mind-blowing.”

The variants are so far neutralized by the vaccine, but Pfizer is working on an updated vaccine to prepare for the day when one could be needed. It picked Beta, the variant with the greatest reduction in virus neutralization in the assay, as the prototype.

“We did not believe we would need a Beta variant vaccine. We did it explicitly as a test case to smooth the pathway in case we do need to do it,” said Dormitzer. Studying the Beta vaccine has taken months, but Pfizer’s goal is to go from synthesis to creating a new vaccine that’s ready to roll out within 100 days.

Achieving this goal largely depends on regulatory oversight and how much testing is required by the Food and Drug Administration; Pfizer’s hope is that the foundation of evidence on existing shots can help new vaccines proceed more quickly. New flu strain vaccines, for example, are slight adjustments from earlier doses and don’t need to be tested in humans before they’re first used. “It’ll take time to get that kind of confidence [for Covid vaccines],” said Dormitzer.

There’s also no definitive test to show when a new vaccine might be needed. Beta’s neutralization response was around a third lower than for the original variant — which sounds significant, but is a relatively small difference. “On the scale of flu viruses, you don’t pay attention until it gets beyond a fourfold reduction in neutralization,” said Dormitzer. And the difference didn’t map onto a change in real life: The original vaccine held up well against Beta in efficacy studies.

Kena Swanson, senior director of viral vaccines. Desiree Rios for STAT

Researchers are still trying to figure out what neutralizing data is a sign for concern. There doesn’t seem to be a linear correlation between the number of  antibodies and level of protection. Instead, there’s likely a threshold, below which we would be vulnerable to severe disease. “There certainly seems to be some trend, but whether there’s an absolute threshold, we still don’t have an understanding,” said Swanson.

Studying Covid-19 has only emphasized how little we know. “A lot of times, when you’re working in this field, you’d look at the animal data and say this thing has a wimpy antibody response, let’s not go ahead with it,” said Roopchand. Data from the Phase 3 efficacy study of the Covid vaccine undermines that approach: Vaccinated participants have protections against the virus by day 12, at a time when there’s barely any antibody response. “That was the biggest surprise,” said Roopchand.

He’s hoping to continue dissecting the data, to try and find other indicators of effective defenses. T cells, which are produced by the immune system to destroy virus-infected cells, and antibody-dependent cellular cytotoxicity (ADCC), a specific immune response whereby a cell is covered in antibodies then destroyed by white blood cells, have largely been neglected in immunology research, said Roopchand. Perhaps you don’t need an overwhelming antibody response, and it’s more important to measure cellular immunity. “The 12-day protection data is telling us there’s more to it,” he said. “This is a great time to learn.”

The work is relentless. Dormitzer said he didn’t receive emails from 9:30 to 11 one night earlier this month. “I actually looked up the help desk, I wanted to see if there was something wrong with my computer. Normally my inbox would be full at that time.”

He’s known as a nocturnal figure at Pfizer, yet seems calm and rested, apparently without a need for sleep. At times of pressure, such as when the neutralization data came in lower for Beta and Pfizer decided it needed to share it publicly as quickly as possible, Shi said he worked until late in the evening, and then Dormitzer took over from 10 p.m. to 5 a.m. “He works the night shift, he has the capability to do that,” he said.

The constant intensity is both rewarding and wearing, especially as large portions of the public reject Pfizer’s lifesaving vaccine. Tompkins, the graphing unicorn, is soft-spoken, but gently hits the table with frustration as she talks about how the pandemic might end. “We’re going to continue until everyone is vaccinated,” she said.

More adaptations are coming. The team is currently waiting on the results for how the vaccine fares against B.1.621, a variant that originated in Colombia that doesn’t yet have a Greek letter name.

The worst-case scenario is for a variant to develop that escapes vaccine protections entirely. Pfizer’s scientists are divided on whether that could happen. “I don’t think we’ll be back to square one,” said Dormitzer. His colleague Roopchand thinks he’s overly optimistic. A “new square one” is still possible, he said.

The pandemic feels interminably long but, from a research perspective, Covid-19 is young. “It’s still early days in understanding how the virus is evolving and understanding what might come next,” said Swanson. New variants are inevitable, only their impact is impossible to predict.

Create a display name to comment

This name will appear with your comment