FluPredictibility

This repository contains the code used in the article

Limited predictability of clade frequencies of seasonal influenza viruses
Pierre Barrat-Charlaix, John Huddleston, Trevor Bedford & Richard Neher

Installing

You need Julia version 1.10 or higher to run this code, and IJulia to run the notebooks.
First, clone this repository to your computer. Navigate to it and start a julia session by typing julia --project=.. Alternatively, run julia and type using Pkg; Pkg.activate(".") to activate the project environment. Finally, run using Pkg; Pkg.instantiate(). This will download all the dependencies given in the Manifest.toml file.

You should now be able to run the notebooks. Still in the julia session, run using IJulia; IJulia.notebook() (this will potentially prompt you to install jupyter through the Conda.jl package, explanations here). Jupyter will start in a browser window, from which you can navigate to the FluPredictibility folder. Double-click any notebook to run it.

Content

Important:

• Folder data contains trees and lists of positions in HA and NA proteins necessary for the analysis. It does not contain the sequence alignments used in the paper. Accession numbers of strain used are available as a supplementary file with the publication.
• Folder notebooks contains jupyter notebooks that generate the figures of the paper. These notebook expect a folder alignments to exist, containing amino acid or nucleotide alignments to be used. Alignments should be put there in the fasta format.