Advancing influenza antigenic analysis to improve vaccine effectiveness

My project aims to help make the flu vaccine more effective.

Flu is a common illness that people catch many times in their lives. Flu can cause severe disease or death, particularly in children, the elderly and people with long-standing health problems. It is recommended that people in these groups get the flu vaccine to avoid getting ill.

Vaccines work by teaching the body to recognise and defeat an infection. Here's the problem: there are many different varieties of flu virus and a different vaccine is needed against each one. Different varieties spread each year, and the difficulty is predicting which flu virus to make the vaccine with.

Currently, scientists and doctors constantly gather and test flu virus samples from around the world. This tells them which flu types might be spreading to help decide what to put in the vaccine. Sometimes the virus type can change after the vaccine choice, and this makes the vaccine work less well.

A mathematical method has been developed that shows how different flu viruses are related as a picture, a kind of map that shows how similar different viruses are. Viruses that are like each other will appear close together in the map and viruses that are more different will be further apart on the map. By seeing where the virus may move to next on this map, we can see which types to target with our vaccines.

There are five events in the past decade that mean I need to update this mathematical method.

1. Changes in some flu viruses mean that a less reliable experimental measurement must be used. The current method was designed for more reliable measurements.

2. There is now a large amount of data that would be more useful if it could be analysed all together. We would understand a lot more about how flu changes if all the information was linked up.

3. There are new differences in the pattern of changes of flu viruses that I want to investigate further. In recent years, there have been more but smaller changes in flu viruses, and old flu viruses are not being replaced so quickly.

4. Traditional methods using animal experiments no longer give helpful results. Instead, blood samples from people need to be used. People have had flu and flu vaccines in the past so their blood samples are more complicated to understand.

5. We can now make changes to viruses in the laboratory to help us understand how viruses change in nature. We now need to understand how natural flu virus types relate to these laboratory changed virus types.

To improve this mathematical method, I will generate practice data, where I know what the outcome would be. I make these maps using the practice data and see how good they are. I will make the practice data similar to real data, with missing, inconsistent or insufficient data! When the current method does not work well with the practice data, I will generate warnings that the results may not be reliable. I will also seek mathematical ways to improve the results.

There are difficulties with experimenting on current flu viruses and scientists need to do different types of experiments to get the necessary information (less reliable experiments and using human blood samples). I will update the mathematical method to be able to cope with these changes. I will make sure that the people who decide which viruses to use in the upcoming vaccines have the best interpretation of their data.

Reliable, up-to-date information on current viruses will also help us to understand the effects of changing viruses in the laboratory.

In view of the large percentage of the population affected by flu, the flu vaccine is very important. Intensive and careful development of new ways of modelling flu virus changes will help choose the best flu vaccine to keep children and the elderly well in winter.