Background

We now have a better understanding of what data are published by Nextstrain's pipeline and how we can use it to get the COVID genome sequences+metadata required to generate target data for the variant nowcast hub.

Currently, the target data pipeline gets Genbank sequences via NCBI's API. It then uses an NBCI command line tool to format the accompanying metadata.

However, Nextstrain also sources Genbank sequences from NCBI, and they publish both the sequences and sequence metadata: https://docs.nextstrain.org/projects/ncov/en/latest/reference/remote_inputs.html#remote-inputs-open-files

Switching to Nextstrain as a source for the target data pipeline information (versus getting it directly from NCBI) would have several advantages:

1. Allow code re-use between the target data pipeline and the process to generate each round's clade list (e.g., using the same filters for host, etc.)
2. Eliminate the need for #1
3. Simplify the code base (reduces issues such as catch and fail early on errors in downloading data #5)