Tracking influenza evolution within individual patients may help predict trends

Science Spotlight

Tracking influenza evolution within individual patients may help predict trends

From the Bloom Laboratory , Basic Sciences Division

Sept. 18, 2017

Influenza infects 3 to 5 million people annually and causes several hundreds of thousands of deaths1.  Each year, public health organizations such as the World Health Organization (WHO) use data to predict which variants of the virus will be included in the annual flu vaccine. Predicting these trends in viral evolution is a crucial public health concern.  In a recent study published in eLife, scientists in the Bloom Laboratory (Basic Sciences Division) analyzed deep sequencing data across long-term flu infections in immune-compromised patients to get a sense for the variability and evolution of the viruses that proliferate within individuals. Interestingly, they identified several mutations that appeared in more than one patient and a pair of mutations that also appeared in most flu viruses that were collected from patients around the globe several years later. 

graphic of a person containing multiple flu viruses

Monitoring the evolution of viruses within individual patients may help to predict global trends.

Image courtesy of Katherine Xue.

The basis of evolution is the spontaneous appearance of mutations that can either be harmful or beneficial to organisms or viruses. These mutations are often the result of errors that occur during the DNA replication process required for reproduction. While many mutations cause problems, some mutations improve the ability of an organism or virus to reproduce. Often the effects of a mutation are only observed over long periods of time, since the impact may be minor and the reproductive cycle of a particular organism may be longer. Viruses, on the other hand, often evolve rapidly due to their constant replication and need to evade their host's immune system long enough to continue reproducing and infecting new host cells. 

While flu infections typically last less than a week, infections in immune-compromised patients can last several weeks to months. Periodic nasal wash samples from these patients present a unique opportunity to observe influenza virus evolution within patients. Researchers in the Bloom Laboratory, led by graduate student Katherine Xue, deep sequenced viral RNA from some of these samples and analyzed the frequency of observing different flu mutations.  "Flu infections are usually short, but evolutionary changes that happen during these short infections sometimes go on to spread around the world. We wondered how much flu viruses could change in a single person while they're sick," said Xue.

Influenza viruses are asexual clones, meaning they are unable to recombine or swap RNA at any random place in their genome when they encounter another flu virus—they can only swap defined segments of their RNA. Because of this, mutations within each segment can only arise in the presence of whatever existing mutations there are in that segment. In this way, two different viruses which each have a unique mutation in the same segment that makes them better at reproducing may appear but these viruses cannot swap sequences within the segment to form an even more capable virus. Thus, determining which new viral sequences become most common based on how beneficial they are in a single or even small group of short infections may not represent which mutations become most common globally. For this reason, a large number of samples from different infections are required.

By monitoring the rise and fall in frequency of different flu virus variants across multiple patients, scientists could potentially get a better sense for which mutations become common in specific genetic backgrounds. For example mutation Y may only become common following mutation X.  Indeed, Xue and her colleagues identified these kinds of trends in their datasets. The scientists specifically analyzed the influenza surface proteins hemagglutinin (HA) and neuraminidase (NA), as mutations are most commonly observed within them. The four patients analyzed were treated with the flu drug oseltamivir and the known drug-resistant mutations T242I and R292K in NA emerged and persisted throughout the infections. Interestingly, the V223I and N225D mutations in HA were found in multiple patients and then became the most commonly observed variant in the global human population at those sites in the years following the infections. Other mutations that became common among multiple patients, however, such as L427F within HA, remained rare globally.

Said Dr. Bloom, "A remarkable aspect of influenza virus's evolution is that every change that spreads around the globe (to hundreds of millions of people each year) must have started in one single person. This study provides insight into how those changes get started within individuals."


Xue KS, Stevens-Ayers T, Campbell AP, Englund JA, Pergam SA, Boeckh M, Bloom JD. "Parallel evolution of influenza across multiple spatiotemporal scales." eLife. 2017;6:e26875 DOI: 10.7554/eLife.26875.


This research was supported by the National Institutes of Health, the Simons Foundation, Howard Hughes Medical Institute, the National Science Foundation, and the Hertz Foundation. 

Competing interests statement: JAE reports research support from Gilead Sciences, GlaxoSmithKline, Chimerix, and Pfizer. MB reports research from Aviragen Therapeutics, Gilead Sciences, Ansun BioPharma, and GlaxoSmithKline. The other authors declare no competing interests.