Combining Influenza hemagglutinin antigenic maps with deep mutational scanning data
Ákos Gellért[1,2] , Oz Kilim[1] , Anikó Mentes[1] and István Csabai[1] (2023.02.15 - 2023.12.15)
[1] ELTE Department of Physics of Complex Systems [2] ELKH Veterinary Medical Research Institute
Kivonat: The first recorded pandemic of the flu occurred in 1580 and since then, flu pandemics have occurred several times throughout history, with the most severe being the Spanish flu in 1918-1919 which killed millions of people worldwide. In the 20th century, significant progress was made in the understanding of the virus and the development of vaccines, which have greatly reduced the impact of flu pandemics. Despite this progress, the flu continues to be a major public health issue, with millions of cases reported each year and an annual death toll in the tens of thousands.
Hemagglutinin, a surface membrane protein of the Influenza virus plays an important role in the infection process of the virus, as it allows the virus to attach to and penetrate host cells. The flu vaccine is formulated each year based on which strains of the virus are predicted to be most prevalent, and it is designed to stimulate the body's immune response to the hemagglutinin protein on those strains. Many antigenic maps have been constructed this far, which reveal the relationships between different strains of a virus, specifically with regards to the way their antigens [1] (e.g., hemagglutinin) are recognized by the immune system. Experimental Influenza HA deep mutational data [2] are also available for the research community to explore the virus functions.
In this project, we aim to in silico combine antigenic maps and deep mutational scanning data to obtain a more comprehensive understanding of the evolution and functional properties of Influenza virus. For example, combining antigenic map data with deep mutational scanning data can provide information about how different mutations affect the ability of a virus to evade the immune response, as well as which regions of the virus are critical for this evasion. This information can be used to inform the design of vaccines and antiviral drugs that target specific regions of the virus that are critical for its function and evolution. We will use AlphaFold2 [3] and ESMFold2 [4] the fastest AI based and most reliable protein structure prediction applications in the world to generate single and/or multiple mutant structures of various Influenza HA protein.
[1] Antigenic map. [2] Flu HA DMS.. [3] J. Jumper et al., “Highly accurate protein structure prediction with AlphaFold,” Nat. 2021 5967873, vol. 596, no. 7873, pp. 583–589, Jul. 2021, doi: 10.1038/s41586-021-03819-2. [4] ESMFold.