Mining evolutionary network of influenza viruses : pipeline for co-mutational analysis of influenza viruses’ sequences

Influenza is a persistent threat to humans, resulting in millions of cases of severe illnesses and about 250,000 to 500,000 deaths. Consequently, it also causes tremendous economic losses as vaccines are produced yearly to counter against influenza. However, a vaccine loses its effectiveness in the following year, since influenza evolves virulently. Therefore, the vaccine must be updated each year to stay effective. Hence, there is a demand to outline the evolution of influenza viruses for better selection of vaccines and predict pandemic strains. In this project, we aim to explore and perfect a novel method that relies on phylogenetic tree which are sampled from sequences data for mutations identification. This method is applied to the hemagglutinin of influenza H3N2. With network analyses integrated in the pipeline, not only does this approach allows us to recover site-pairs that potentially co-mutate, we are able to extract the direction of the relationship as well. From these recovery, we are able to facilitate the design of vaccines to better combat the threat of influenza viruses. We benchmarked the proposed method against a simulated set of data to identify the significant co-mutational site-pairs. The simulated data is built by simulating the substitution processes along a fixed phylogeny with a given nucleotide substitution model and suitable parameters. The results show that the proposed pipeline is robust and is capable of identifying significant dominant mutations during evolution.Bachelor of Engineering (Computer Science