How flu evolution drives outbreak severity

Worldwide, there are about a billion cases of flu every year. Most of them occur during the winter in each geography, with the virus making a huge continuous lap of the globe and evolving as it goes. It’s this gentle genetic “drift” that means the flu can always present a fresh face to our immune systems each year, and keep on infecting us throughout our lives. We combat this by periodically updating our vaccines to better reflect the circulating strains. Traditionally, scientists have gauged the likely effectiveness of those vaccines by looking at the antibodies they provoke; they also keep tabs on the underlying genetic changes that the virus is accruing. But which, if any, is the better guide, and can we improve on it, perhaps by considering both elements side by side? Speaking with Chris Smith, Amanda Perofsky is at the University of Washington…

Amanda - So traditionally the antibody data and the genetic sequences have been analysed separately, just totally in isolation. Scientists look at genetic mutations in the flu viruses and then separately they look at data from the antibody experiments. But now there's more recent methods where we can combine both of those two different sources of information on flu evolution into one analysis. And the analysis is that we create family trees with the flu viruses that aren't just informed by the genetic sequence data, but also by the antibody data.

Chris - And how do you then work out how severe an impact that particular combination will have?

Amanda - So we have historical data on flu outbreaks in the US, and this data is publicly available provided by the US Centers for Disease Control and Prevention. And so there's about 20 years of historical data on people going to the doctor for flu-like symptoms and how those doctor's visits for flu-like symptoms compare to sort of like the overall rate of people going to the doctor. So these are the number of people going to the doctor for flu versus everyone going to the doctor. We can get a rate of flu-like illness over the years and we can break that down by geography as well. We look at each flu season, what was the size of the epidemic, when did the epidemic peak, how severe was the flu season? There's also data collected on deaths that are potentially attributable, which age groups are being infected more in a particular season. And from those family trees of the flu viruses, we can look at how much has the population of flu viruses as a whole changed compared to the last season. And so we get like a one-to-One comparison of this is how much the flu virus population has evolved compared to different aspects of the flu each season.

Chris - And what does that tell you when you've done that? What trends or what pattern emerges in terms of being able to make a better prediction regarding how, how bad a flu season's likely to be based on what the virus is doing?

Amanda - So what we found matches up with the long-held expectations concerning evolution of the flu viruses causing people to become more susceptible to the flu. And in that I mean that the viruses are so different than what you've seen your body or your immune system has seen before, that they're able to infect you even though you've been infected in the past or vaccinated in the past. And so we found further evidence that when flu viruses evolve, that causes bigger flu epidemics. But a thing that was unique about our study is that we looked at all sorts of different ways to measure flu evolution. I mentioned that you can measure flu evolution using the antibody data or the genetic data. What we found that was sort of surprising was that changes in flu genetic sequences were actually more predictive of different aspects of flu epidemics than the antibody data. Even though the antibody data has been considered the gold standard for measuring changes in flu evolution over time

Chris - Is the implication of that then that what we should be doing in future is when we collect samples of flu from patients, we should be looking at how it is evolving and using that evolutionary data to make predictions about vaccine effectiveness about the scale of a likely forthcoming outbreak, rather than assuming, oh look, we've got a reasonable match between the circulating virus and what's in the vaccine, so we're probably gonna be okay...

Amanda - Scientists are already looking at changes in genetic sequences. It's just that perhaps the genetic sequence data should be considered at the same level this gold standard data has been considered in, in terms of determining how much flu viruses have evolved since the last season. Another finding I didn't mention is that traditionally scientists have looked at evolution in one particular gene of the flu virus that is considered to drive the majority of the flu virus escaping detection by immune system. But there's another gene as well we looked at that isn't currently really considered in surveillance of flu evolution. And so we found that this other gene also has just as strong of a relationship with different aspects of flu outbreaks as the gene that scientists have traditionally looked at in the past.