Last fall, a Fred Hutch Cancer Center graduate student hopped on a 4 a.m. videoconference call with scientists from the World Health Organization to talk about a blood-testing method developed at Fred Hutch that could improve the effectiveness of seasonal flu vaccines.
Caroline Kikawa, a graduate student in the University of Washington Medical Scientist Training Program, working with Fred Hutch evolutionary biologist Jesse Bloom, PhD, was excited to talk about the method the Bloom Lab has developed, but she was also nervous.
“That was the first time that we did this at scale for this purpose,” Kikawa said. “It was really stressful because I didn't have any of the data yet.”
She had done much of the preliminary work using the method, which provides a near real-time picture of how well the human immune system is coping with current strains of seasonal influenza, but she didn’t have results ready to share.
“We will have this data for you in a month,” Kikawa told the scientists from the WHO serological assay team on the call.
Kikawa made good on her promise, producing more than 25,000 measurements comprising the largest-ever single-study dataset of influenza-fighting antibodies, which was recently published in the journal Virus Evolution.
She completed almost all the lab work herself in under six months and immediately made her results publicly available to the scientific community, which was noted when she won a Beyond The Journal award for exemplary data sharing.
Her data helped inform the recommendations made in the fall of 2025 about the composition of the flu vaccine that folks from Cape Town to Sydney to Buenos Aires are now receiving in 2026 as autumn begins south of the equator.
And earlier this year, she provided updated data that helped inform composition recommendations about the flu vaccine that Seattle and the rest of the Northern Hemisphere will receive next fall.
The seasonal influenza virus constantly evolves to evade the immunity we build up from past infections and vaccinations, which is why annual vaccinations, tweaked to reflect new strains, are recommended.
Manufacturers require several months to prepare vaccines for the upcoming winter flu season using standard methods, which means they must predict which strains are likely to be the most dominant eight to 12 months in advance.
Researchers consider both genetic and blood test data to get a clear picture of how the virus is adapting all over the world, with each hemisphere getting a preview by observing what’s going on with the virus during the other hemisphere’s flu season.
Genetic data shows which virus strains are most common, and blood test data shows how well the immune system’s antibodies are fighting those strains. The changing genetics of the virus help scientists make better sense of our changing immunity and vice versa.
The WHO convenes two meetings each year to integrate those observations and determine if the strains in circulation have changed enough to warrant an update to the vaccine.
The WHO makes recommendations in September for the vaccine that will be administered in the Southern Hemisphere’s next flu season, and again in February for the vaccine the Northern Hemisphere will receive later in the year during its flu season.
The molecular component of the virus that evolves the fastest is a protein called hemagglutinin (HA), which is the H in the H1N1 and H3N2 influenza A flu viruses, which the WHO tracks along with influenza B.
Though the WHO routinely collects blood samples from patients at hospitals all over the world, researchers typically test flu strains against the immune systems of ferrets that have been inoculated with one of many available vaccine candidate viruses. Ferrets share enough of our biology and lung structure to stand in for humans. When ferret antibodies to the current vaccine virus no longer bind to recently circulating viruses, that’s a signal that we may need a new vaccine. Successful binding of ferret antibodies to new vaccine candidates tells researchers which candidates might be most effective.
“You can take a virus that infects humans, you can put it into ferrets, and there's no changes that need to happen for the virus to successfully infect that ferret,” said John Huddleston, PhD, a staff scientist in the lab of computational biologist Trevor Bedford, PhD, who specializes in using genetics to provide accurate forecasts of how the flu virus is likely to evolve from season to season. Bedford is a Howard Hughes Medical Institute Investigator.
But there are drawbacks to using ferrets, which are usually exposed to only one specific flu strain, unlike humans who have a long history of multiple exposures to multiple strains as well as annual flu shots to boost immunity.
Testing human blood makes more sense, but there’s too many different strains of flu and not enough blood samples to test one strain against one blood sample the way it’s done with ferrets.
In the last few years, however, the Bloom Lab, in the Basic Sciences Division of Fred Hutch, found a way to make more efficient use of the limited supply and make it work at a scale large enough to update the seasonal flu vaccine.
A few years ago, Andrea Loes, PhD, a staff scientist and lab manager in the Bloom Lab, developed a new way to rapidly measure in a single experiment how the antibodies in human blood fight many different flu strains.
Federal Funding Helps Us Save Lives
At Fred Hutch, 70% of our research funding comes from federal grants, which are awarded based on scientific merit. This level of federal funding is a direct reflection of our proven ability to make landmark breakthroughs and our trusted ability to lead large scientific collaborations.
Traditional methods only test a single blood sample against a single viral strain, which is effective if researchers pick the right strain or couple of strains to analyze, hoping they’re representative of what’s circulating or will be circulating in eight months to a year.
But if they choose the wrong ones, the vaccine will miss the mutations that help the virus hide from the immune system.
The test Loes developed for the Bloom Lab basically followed the same workflow as the traditional method, but instead of testing one strain per sample, it tested all the strains at once.
Here’s how they do it:
The key differences in the flu virus they want to track involve mutations to the HA protein, so for each strain they add a unique nucleotide barcode tacked onto the HA gene that acts like a nametag.
The barcoded HA gene is then stitched into the backbone of a lab-adapted flu strain commonly used for experiments under safe laboratory conditions.
Then the next step is to pool all those barcoded virus strains into a library that can be tested against a blood sample all at once. Every strain in the library is identical except for the different, barcoded HA gene.
They apply the library to the sample and then the mixture stews as the virus strains attempt to infect the cells. If the antibodies can't block the strains, the infected cells make more RNA with that barcode.
After 16 hours, it’s time to extract whatever viral RNA has been produced from the infected cells.
But first, they add a fixed amount of RNA that has nothing to do with the viruses, which serves as a common reference to measure the viral RNA against.
This is required because viral RNA amounts may vary because of technical factors of the experiment itself that affect extraction and distort the results if they aren’t accounted for in the analysis.
By adding control RNA, which is also barcoded, to the mix, they can measure the amount of viral RNA relative to the fixed amount of non-viral RNA that undergoes the same sequencing process.
As long as both the viral RNA and the fixed amount move up or down in the same ratio, they know that technical differences aren’t muddying the signal.
The sequencing data reveals which strains were knocked out by antibodies in the blood sample (they didn’t generate any viral RNA) and which ones produced the strongest infections (they produced the most viral RNA, relative to the control amount).
In the last few years, the lab has tested the method on “proof-of-concept” experiments, but never at the scale required to produce useful data for updating the seasonal flu vaccine for an entire hemisphere.
Kikawa began the University of Washington MD PhD program in 2020 and joined the Bloom Lab in 2022.
She spent her first year working on Zika virus with co-mentor Leslie Goo, PhD, MPH in the Vaccines and Infectious Diseases Division, but when Goo left Fred Hutch to work in industry, Kikawa had to figure out what to do next.
“I had this pivot moment with Jesse,” she said.
Bloom, a Howard Hughes Medical Institute Investigator, told her about Loes’ work on the new blood test method for flu, which Kikawa knew about from presentations in lab meetings.
“He said, ‘I think this is exciting,’ and I said ‘I know, I think this is exciting too,’ so I thought, OK, I'll pivot to flu,” Kikawa said.
In 2025, she and her colleagues launched an effort to generate data with the new method that could be ready in time for the WHO meeting in September that would recommend updates to the flu vaccine for the Southern Hemisphere.
That began in May with the design of the library of 140 influenza A strains representing the array of human H3N2 and H1N1 viruses circulating from April to May of 2025.
“These flu viruses are making mistakes all the time as they make more copies of themselves,” Huddleston said. “And those mistakes are the mutations that we talk about, but they can be completely random. The ones that are successful are usually the ones that escape our immunity. And so, we have to figure out — out of all the things that have happened — which ones are really important.”
That’s where the Bloom Lab method comes in, which tests a component of blood called serum, the liquid part minus the blood cells and clotting proteins.
“The serological data helps you understand, try to understand what the changes in the sequence data actually matter for the virus to escape immunity,” Huddleston said.
By the time Kikawa and Bloom had to share her work on that 4 a.m. videoconference call last fall, she was well on her way.
She had completed work on the library and was ready to test it against 188 human serum samples — collected between October 2024 and April 2025 — that represented a wide range of ages and geographic locations, including Hong Kong and Seattle.
A few weeks later when the results were ready to share, the Bloom Lab posted them for the entire scientific community, including the WHO.
The Bloom Lab points out that its results are made publicly available for anyone to use, including other scientists, public health agencies, biotech companies and vaccine manufacturers. The anyone includes the WHO — a 194-member global health alliance that the United States helped establish in 1948.
Although the U.S. was the organization’s most influential member and largest financial contributor, President Trump announced soon after taking office for his second term that the U.S would withdraw its membership and support.
The U.S. formally departed in January of 2026, though U.S. scientists continue to participate in different ways, including on decisions about the composition of seasonal flu vaccines.
It’s too soon to know whether Kikawa’s data will improve the efficacy of the flu vaccine the Southern Hemisphere is now receiving, but her method picked up some important nuances that would have been missed by just testing ferret blood the traditional way.
In the H1N1 subtype, for example, ferret blood didn’t show any changes that would help the virus escape our immunity, even though it’s clear from the Northern Hemisphere’s experience with flu this winter that some changes resulted in greater immune escape.
The team worked with Huddleston and Bedford to map their immunity results onto interactive family trees of the flu virus using Nextstrain, an open-source website Bedford co-founded in 2015 to help researchers visualize the genomic evolution of influenza and other global pathogens, such as SARS-CoV-2.
“So that just tells us that the ferret readout is not telling us the full picture,” Huddleston said. “And then when we look at Caroline's data for H1N1, we see a clear separation of recent H1N1 groups.”
Her data predicted which of those changes was more likely to succeed and data from the last few months in the Northern Hemisphere proved her right, which means the recipe for the Southern Hemisphere’s vaccine stands a better chance of success.
Kikawa will return to medical school in March 2027 after completing her PhD, and anticipates graduating with her MD in 2029. She’s learned a lot since her pivot to the flu.
“It’s been so amazing to work with John, Andrea and Jesse really closely on this,” she said. “It was a little bit of a stressful way to start a PhD, but it's been really exciting and feels like it has the potential to be meaningful.”
This work is funded by grants from the National Institutes of Health, UK Medical Research Council, Grants-in-Aid for Emerging and Reemerging Infectious Diseases from the Ministry of Health, Labour and Welfare, Japan, Dolores Covarrubias and the Genomics & Bioinformatics Shared Resource of the Fred Hutch/University of Washington/Seattle Children’s Hospital Cancer Consortium and by Fred Hutch Scientific Computing and Research Grants Council from the University Grants Committee of Hong Kong.
John Higgins, a staff writer at Fred Hutch Cancer Center, was an education reporter at The Seattle Times and the Akron Beacon Journal. He was a Knight Science Journalism Fellow at MIT, where he studied the emerging science of teaching. Reach him at jhiggin2@fredhutch.org or @jhigginswriter.bsky.social.
Are you interested in reprinting or republishing this story? Be our guest! We want to help connect people with the information they need. We just ask that you link back to the original article, preserve the author’s byline and refrain from making edits that alter the original context. Questions? Email us at communications@fredhutch.org
Every dollar counts. Please support lifesaving research today.
For the Media